vcs bundled agents 601 lin

8/12/2019 Vcs Bundled Agents 601 Lin

http://slidepdf.com/reader/full/vcs-bundled-agents-601-lin 1/279

Veritas ™ Cluster Server 6.0.1

Bundled Agents ReferenceGuide - Linux

September 2012



Symantec Corporation350 Ellis StreetMountain View, CA 94043

http://www.symantec.com

http://www.symantec.com/

http://www.symantec.com/



Technical SupportSymantec Technical Support maintains support centers globally. TechnicalSupport ’ s primary role is to respond to specific queries about product featuresandfunctionality. TheTechnical Support group also createscontent forouronlineKnowledge Base. The Technical Support group works collaboratively with theother functional areas within Symantec to answer your questions in a timelyfashion.Forexample, theTechnicalSupport groupworkswith Product EngineeringandSymantec Security Response to provide alerting services andvirus definitionupdates.

Symantec ’ s support offerings include the following:

■ A range of support options that give you the flexibility to select the right

amount of service for any size organization■ Telephone and/or Web-based support that provides rapid response and

up-to-the-minute information

■ Upgrade assurance that delivers software upgrades

■ Global support purchased on a regional business hours or 24 hours a day, 7days a week basis

■ Premium service offerings that include Account Management Services

For information about Symantec ’ s support offerings, you can visit our Web siteat the following URL:

www.symantec.com/business/support/index.jsp

All support services will be delivered in accordance with your support agreementand the then-current enterprise technical support policy.

Contacting Technical SupportCustomers with a current support agreement may access Technical Supportinformation at the following URL:

www.symantec.com/business/support/contact_techsupp_static.jsp

Before contacting Technical Support, make sure you have satisfied the systemrequirements that are listed in your product documentation. Also, you should beat thecomputer on which theproblem occurred, in case it is necessaryto replicatethe problem.

When you contact Technical Support, please have the following informationavailable:

■ Product release level

http://www.symantec.com/business/support/index.jsp

http://www.symantec.com/business/support/contact_techsupp_static.jsp

http://www.symantec.com/business/support/contact_techsupp_static.jsp

http://www.symantec.com/business/support/index.jsp



Support agreement resourcesIf you want to contact Symantec regarding an existing support agreement, pleasecontact the support agreement administration team for your region as follows:

[email protected] and Japan

[email protected], Middle-East, and Africa

[email protected] America and Latin America

DocumentationProduct guides are available on the media in PDF format. Make sure that you areusing the current version of the documentation. The document version appearson page 2 of each guide. The latest product documentation is available on theSymantec Web site.

https://sort.symantec.com/documents

Your feedback on product documentation is important to us. Send suggestionsfor improvements and reports on errors or omissions. Include the title anddocument version (located on the second page), and chapter and section titles of the text on which you are reporting. Send feedback to:

[email protected]

For information regarding the latest HOWTO articles, documentation updates,

or to ask a question regarding product documentation, visit the Storage andClustering Documentation forum on Symantec Connect.

https://www-secure.symantec.com/connect/storage-management/forums/storage-and-clustering-documentation

About Symantec ConnectSymantec Connect is the peer-to-peer technical community site for Symantec ’ senterprise customers. Participantscanconnect andshare information with otherproduct users, including creating forum posts, articles, videos, downloads, blogsand suggesting ideas, as well as interact with Symantec product teams and

Technical Support. Content is rated by the community, and members receivereward points for their contributions.

http://www.symantec.com/connect/storage-management

mailto:[email protected]


















Technical Support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . 4

Chapter 1 Introducing Bundled agents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . 17

About Bundled agents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17Resources and their attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 18Modifying agents and their resources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . 18

Attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . 18IMF aware agents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19Enabling debug log messages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . 20

Chapter 2 Storage agents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

About the storage agents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . 21DiskGroup agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

Dependencies for DiskGroup agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . 22Agent functions for DiskGroup agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . 22State definitions for DiskGroup agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . 24

Attributes for DiskGroup agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . 24Resource type definition for DiskGroup agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30Notes for DiskGroup agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . 30Sample configurations for DiskGroup agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32Debug log levels for DiskGroup agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . 32

DiskGroupSnap agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32Dependencies for DiskGroupSnap agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . 33Agent functions for DiskGroupSnap agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34State definitions for DiskGroupSnap agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34Attributes for DiskGroupSnap agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . 34Notes for DiskGroupSnap agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . 36Resource type definition for DiskGroupSnap agent . . . . . . . . . . . . . . . . . . . . . . . 40Sample configurations for DiskGroupSnap agent . . . . . . . . . . . . . . . . . . . . . . . . . 40Debug log levels for DiskGroupSnap agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47

DiskReservation agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . 47Agent functions for DiskReservation agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47State definitions for DiskReservation agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47Attributes for DiskReservation agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . 48Resource type definition for DiskReservation agent . . . . . . . . . . . . . . . . . . . . . . 49

Contents



Notes for DiskReservation agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . 49Sample configurations for DiskReservation agent . . . . . . . . . . . . . . . . . . . . . . . . 50Debug log levels for DiskReservation agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51

Volume agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . 51Dependencies for Volume agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . 51Agent functions for Volume agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . 52State definitions for Volume agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . 52Attributes for Volume agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . 53Resource type definition for Volume agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53Sample configuration for Volume agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53Debug log levels for Volume agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . 54

VolumeSet agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54Dependencies for VolumeSet agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . 54Agent functions for VolumeSet agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . 54State definitions for VolumeSet agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . 55Attributes for VolumeSet agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . 55Resource type definition for VolumeSet agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55Sample configurations for VolumeSet agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56Agent notes for VolumeSet agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . 56Inaccessiblevolumes prevent theVolumeSetagentfromcoming

online . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56Debug log levels for VolumeSet agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . 56

LVMLogicalVolume agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56Dependencies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57

Agent functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57State definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57Attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 58Resource type definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . 58Sample configurations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . 58Debug log levels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59

LVMVolumeGroup agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . 59Dependencies for LVMVolumeGroup agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60Agent functions for LVMVolumeGroup agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60State definitions for LVMVolumeGroup agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61Attributes for LVMVolumeGroup agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . 61Resource type definition for LVMVolumeGroup agent . . . . . . . . . . . . . . . . . . 61LVMVolumeGroup agent notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . 62Enabling volume groupactivationprotectionforVeritas Dynamic

Multi-Pathing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . 62Notes for volume group activation protection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62Sample configurations for LVMVolumeGroup agent . . . . . . . . . . . . . . . . . . . . . 63Debug log levels for LVMVolumeGroup agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64

Mount agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . 65

Contents8



Dependencies for Mount agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . 65Agent functions for Mount agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . 65State definitions for Mount agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . 67

Attributes for Mount agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . 68Resource type definition for Mount agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72Notes for Mount agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . 73Support for spaces in directory names . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . 74Support for multiple bindfs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . 74High availability fire drill . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . 75VxFS file system lock . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . 75IMF usage notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76Enabling Level two monitoring for the Mount agent . . . . . . . . . . . . . . . . . . . . . 76Sample configurations for Mount agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . 77Debug log levels for Mount agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . 77

Chapter 3 Network agents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79

About the network agents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . 79Agent comparisons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . 79

IP agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . 80High availability fire drill for IP agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 81Dependencies for IP agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 81Agent functions for IP agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . 81State definitions for IP agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . 82Attributes for IP agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . 82Resource type definition for IP agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . 85Sample configurations for IP agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . 86Debug log levels for IP agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . 86

NIC agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . 86High availability fire drill for NIC agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . 87Dependencies for NIC agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . 87Bonded network interfaces for NIC agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87Agent functions for NIC agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . 88State definitions for NIC agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . 88Attributes for NIC agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 89

Resource type definition for NIC agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . 91Monitoring bonded NICs for NIC agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . 91Setting Mii and miimon for NIC agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . 91Case 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . 91Case 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . 92Case 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . 92Sample configurations for NIC agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . 92Debug log levels for NIC agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . 93

Contents



IPMultiNIC agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93Dependencies for IPMultiNIC agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . 93Agent functions for IPMultiNIC agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . 94

State definitions for IPMultiNIC agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . 94Attributes for IPMultiNIC agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . 95Resource type definition for IPMultiNIC agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96Sample configuration: IPMultiNIC and MultiNICA .. . . . . . . . . . . . . . . . . . . . . . . 97Debug log levels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100

MultiNICA agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100Dependencies for MultiNICA agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . 101IP Conservation Mode (ICM) for MultiNICA agent . . . . . . . . . . . . . . . . . . . . . . . 101Performance Mode (PM) for MultiNICA agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102Agent function for MultiNICA agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . 102Attributes for MultiNICA agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . 103Resource type definition for MultiNICA agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109Sample configurations for MultiNICA agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109IPv6 configuration for MultiNICA agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 112Mixed mode configuration — IPv4 and IPv6 for MultiNICA

agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 113Debug log levels for MultiNICA agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . 114

DNS agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . 114Dependencies for DNS agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . 115Agent functions for DNS agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . 115State definitions for DNS agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . 116

Attributes for DNS agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . 117Resource type definition for DNS agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123Agent notes for DNS agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . 123Sample configurations for DNS agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . 128Debug log levels for DNS agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . 130

Chapter 4 File share agents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131

About the file service agents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . 131NFS agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . 131

Dependencies for NFS agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . 132

Agent functions for NFS agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . 132State definitions for NFS agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . 133Attributes for NFS agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . 133Resource type definition for NFS agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . 134Notes for NFS agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134Sample configurations for NFS agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . 135Debug log levels for NFS agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . 135

NFSRestart agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136

Contents10



Dependencies for NFSRestart agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . 136Agent functions for NFSRestart agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . 137State definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . 138

Attributes for NFSRestart agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . 139Resource type definition for NFSRestart agent . . . . . . . . . . . . . . . . . . . . . . . . . . . 139Notes for NFSRestart agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . 140Sample configurations for NFSRestart agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141Debug log levels for NFSRestart agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . 142

Share agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . 142Dependencies for Share agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . 142Agent functions for Share agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . 143State definitions for Share agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . 143Attributes for Share agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . 144Resource type definition for Share agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145Notes for Share agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . 145Sample configurations for Share agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . 146Debug log levels for Share agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . 146

About the Samba agents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . 146The Samba agents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . 147Before using the Samba agents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . 147Supported versions for Samba agents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . 147Notes for configuring the Samba agents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147

SambaServer agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149Dependencies for SambaServer agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . 149

Agent functions for SambaServer agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . 149State definitions for SambaServer agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150Attributes for SambaServer agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . 150Resource type definitions for SambaServer agent . . . . . . . . . . . . . . . . . . . . . . . 152Sample configurations for SambaServer agent . . . . . . . . . . . . . . . . . . . . . . . . . . . 152Debug log levels for SambaServer agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152

SambaShare agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153Dependencies for SambaShare agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . 153Agent functions for SambaShare agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . 153State definitions for SambaShare agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153Attributes for SambaShare agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . 154Resource type definition for SambaShare agent . . . . . . . . . . . . . . . . . . . . . . . . . 154Sample configuration for SambaShare agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154Debug log levels for SambaShare agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155

NetBios agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155Dependencies for NetBios agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . 155Agent functions for NetBios agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . 156State definitions for NetBios agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . 156Attributes for NetBios agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . 157

Contents



Resource type definition for NetBios agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158Sample configuration for NetBios agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . 158Debug log levels for NetBios agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . 159

Chapter 5 Service and application agents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161

About the services and applications agents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . 161Apache HTTP server agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . 161

Dependencies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162Agent functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162State definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . 163Attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . 164Resource type definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 167Apache HTTP server notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . 168

Sample configurations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . 171Application agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176High availability fire drill for Application agent . . . . . . . . . . . . . . . . . . . . . . . . . 176Dependencies for Application agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . 177Agent functions for Application agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . 177State definitions for Application agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179Attributes for Application agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . 180Resource type definition for Application agent . . . . . . . . . . . . . . . . . . . . . . . . . . 184Notes for Application agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . 184Sample configurations for Application agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186Debug log levels for Application agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . 187

CoordPoint agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187Coordination Point server as a coordination point . . . . . . . . . . . . . . . . . . . . . . 188SCSI-3 based disk as a coordination point . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188Dependencies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188Agent functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189State definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . 189Attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . 190Resource type definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 190Notes for the CoordPoint agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . 191Sample configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . 192

Debug log levels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192KVMGuest agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192

Dependencies for KVMGuest agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . 193Agent functions for KVMGuest agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 193State definitions for KVMGuest agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . 194Attributes for KVMGuest agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . 195Resource type definition for KVMGuest agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197Notes for KVMGuest agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . 197

Contents12



Sample configurations for KVMGuest agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 198Sample configurations for RHEV environment . . . . . . . . . . . . . . . . . . . . . . . . . . . 201Sample Configuration for SuSE KVM ... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . 202

Debug log levels for KVMGuest agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 203Process agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . 203High availability fire drill for Process agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203Dependencies for Process agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . 204Agent functions for Process agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . 204State definitions for Process agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . 205Attributes for Process agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . 205Resource type definition for Process agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 206Usage notes for Process agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . 207Sample configurations for Process agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207Debug log levels for Process agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . 207

ProcessOnOnly agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207Dependencies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 208Agent functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . 208State definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . 208Attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 208Resource type definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . 210ProcessOnOnly agent usage notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . 210Sample configurations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . 210Debug log levels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 210

Chapter 6 Infrastructure and support agents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213About the infrastructure and support agents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213NotifierMngr agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213

Dependency .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 214Agent functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . 214State definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . 214Attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215Resource type definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . 217Sample configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . 218Debug log levels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219

Proxy agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . 220Dependencies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220Agent functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . 220Attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 221Resource type definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . 221Sample configurations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . 221Debug log levels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223

Phantom agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . 223

Contents



Dependencies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223Agent functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224Resource type definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 224

Sample configurations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . 224RemoteGroup agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225Dependency .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225Agent functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 226State definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . 226Attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . 227Resource type definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 232Debug log levels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 232

Chapter 7 Testing agents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233

About the testing agents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233ElifNone agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 233Dependencies for ElifNone agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . 233Agent function for ElifNone agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . 234State definitions for ElifNone agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . 234Attributes for ElifNone agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . 234Resource type definition for ElifNone agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 235Sample configuration for ElifNone agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 235Debug log levels for ElifNone agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . 235

FileNone agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 235Dependencies for FileNone agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . 235Agent functions for FileNone agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . 236State definitions for FileNone agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . 236Attribute for FileNone agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . 236Resource type definition for FileNone agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 236Sample configuration for FileNone agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 236Debug log levels for FileNone agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . 237

FileOnOff agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . 237Dependencies for FileOnOff agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . 237Agent functions for FileOnOff agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . 237State definitions for FileOnOff agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . 238

Attribute for FileOnOff agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . 238Resource type definition for FileOnOff agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 238Sample configuration for FileOnOff agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 238Debug log levels for FileOnOff agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . 238

FileOnOnly agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 239Dependencies for FileOnOnly agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . 239Agent functions for FileOnOnly agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 239State definitions for FileOnOnly agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 239

Contents14



Attribute for FileOnOnly agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . 240Resource type definition for FileOnOnly agent . . . . . . . . . . . . . . . . . . . . . . . . . . . 240Sample configuration for FileOnOnly agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 240

Debug log levels for FileOnOnly agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 240

Chapter 8 Replication agents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . 241

About the replication agents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . 241RVG agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . 241

Dependencies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 242Agent functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . 243State definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . 243Attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 243Resource type definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . 244

Sample configurations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . 244RVGPrimary agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245Dependencies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245Agent functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . 246State definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . 247Attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 247Resource type definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . 250Sample configurations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . 251

RVGSnapshot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251Dependencies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 252Agent functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . 252State definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . 252Attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 253Resource type definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . 253Sample configurations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . 254

RVGShared agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 254Dependencies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 254Agent functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . 255State definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . 255Attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 256Resource type definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . 256

Sample configurations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . 256RVGLogowner agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . 256

Dependencies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 257Agent functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . 257State definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . 258Attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 258Resource type definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . 259RVGLogowner agent notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . 259

Contents



Sample configurations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . 260RVGSharedPri agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 260

Dependencies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 261

Agent functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 261State definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . 262Attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . 262Resource type definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . 263Sample configurations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . 263

VFRJob agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . 263Overview .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . 263Dependencies for VFRJob agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . 264High availability of scheduler and replicator daemons . . . . . . . . . . . . . . . . . 266Agent functions for VFRJob agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . 266State definitions for VFRJob agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . 267Attributes for VFRJob Agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . 267Resource type definitions for VFRJob agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 267High availability of VFR daemons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . 268Configuration of VFRJob service groups on the source

system .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . 268Sample configuration of VFRJob agent on source system .. . . . . . . . . . . . 268Configuration for VFRJob service groups on the target

system .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . 269Sample configuration of VFRJob agent on target system .. . . . . . . . . . . . . 270Changing file replication direction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . 271

Notes for the VFRJob agent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . 272

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 273

Contents16



Introducing Bundled agents

This chapter includes the following topics:

■ About Bundled agents■ Resources and their attributes

■ Modifying agents and their resources

■ Attributes

■ IMF aware agents

■ Enabling debug log messages

About Bundled agentsBundled agentsareVeritas Cluster Server (VCS)processes that manageresourcesof predefined resource types according to commands received from the VCSengine, HAD. You install these agents when you install VCS.

A node has one agent per resource type that monitors all resources of that type.For example, a single IP agent manages all IP resources.

When the agent starts, it obtains the necessary configuration information fromVCS. The agent then periodically monitors the resources, and updates VCS withthe resource status.

Agents can:

■ Bring resources online.

■ Take resources offline.

■ Monitor resources and report state changes.

For a more detailed overview of how agents work, refer to the Veritas Cluster Server Administrator ’ s Guide .

1Chapter



Resources and their attributesResources are parts of a system. They are known by their types, for example: a

volume, a disk group, or an IP address. VCS includes a set of resource types.Different attributes define these resource types in the types.cf file. Each type hasa corresponding agent that controls the resource.

TheVCSconfiguration file, main.cf,contains thevalues for the resource attributesand has an include directive to the types.cf file.

An attribute ’ s given value configures the resource to function in a specific way.By modifying the value of a resource attribute, you can change the way the VCSagent manages the resource.Forexample, the IP agent uses theAddress attributeto determine the IP address to monitor.

Modifying agents and their resourcesUse the Cluster Manager (Java Console), Veritas Operations Manager, or thecommand line to dynamicallymodify theconfiguration of the resourcesmanagedby an agent.

VCS enables you to edit the main.cf file directly. To implement these changes,make sure to restart VCS.

See the Veritas Cluster Server Administrator ’ s Guide for instructions on how tocomplete these tasks.

AttributesAttributes contain data about the cluster, systems, service groups, resources,resource types, and the agent. An attribute has a definitionand a value. You changeattribute values to configure VCS resources. Attributes are either optional orrequired, although sometimes attributes that are optional in one configurationmight be required in other configurations. Many optional attributes havepredefined or default values, which you should change as required.

A variety of internal use only attributes also exist. Do not modify theseattributes — modifying them can lead to significant problems for your clusters.

Attributes have type and dimension. Some attribute values can accept numbers,others can accept alphanumeric values or groups of alphanumeric values, whileothers are simple boolean on/off values.

Introducing Bundled agentsResources and their attributes

18



Table 1-1 Attribute data types

DescriptionData Type

Enclose strings, which are a sequence of characters, in double quotes("). Optionally enclose strings in quoteswhen they begin with a letter,and contains only letters, numbers, dashes (-), and underscores (_).

A stringcan contain doublequotes,but thequotes must be immediatelypreceded by a backslash. In a string, represent a backslash with twobackslashes (\\).

string

Signed integer constants are a sequence of digits from 0 to 9. You canprecede them with a dash. They are base 10. Integers cannot exceedthe value of a 32-bit signed integer: 2147483647.

integer

A boolean is an integer with thepossible values of 0 (false) and 1 (true).boolean

Table 1-2 Attribute dimensions

DescriptionDimension

A scalar has only one value. This is the default dimension.scalar

A vector is an ordered list of values. Each value is indexed using apositive integer beginningwith zero.A setof brackets([]) denotesthatthedimension isa vector. Find thespecified brackets after theattributename on the attribute definition in the types.cf file.

vector

A keylist is an unordered list of unique strings.keylist

An association is an unorderedlist of name-value pairs.An equal signseparates each pair. A set of braces ({}) denotes that an attribute is anassociation. Braces are specified after the attribute name on theattribute definition in the types.cf file, for example: strSnmpConsoles{}.

association

IMF aware agents■ Application agent. See “ Application agent ” on page 176.■ DiskGroup agent. See “ DiskGroup agent ” on page 21.

■ Mount agent. See “ Mount agent ” on page 65.

■ Process agent. See “ Process agent ” on page 203.

Introducing Bundled agentsIMF aware agents



Enabling debug log messagesTo help troubleshoot agent issues, youcanenabledebug log messages in theagent

framework as well as the agents.To enable agent framework debug log messages:

# hatype -modify agent_name LogDbg -add DBG_AGDEBUG DBG_AGINFODBG_AGTRACE

For example:

# hatype -modify Mount LogDbg -add DBG_AGDEBUG DBG_AGINFO DBG_AGTRACE

To enable agent-specific debug log messages:

# hatype -modify agent_name LogDbg -add debug_log_levels

For example:

# hatype -modify Mount LogDbg -add DBG_1 DBG_2 DBG_3 DBG_4 DBG_5 DBG_6

Alternatively, you can also use the following command:

# hatype -modify Mount LogDbg -add 1 2 3 4 5 6

Agent-specific debug log level information is specified in the agent ’ s description.For example, for information about the Mount agent, See “ Debug log levels for

Mount agent”

on page 77.

Introducing Bundled agentsEnabling debug log messages

20



Storage agents


■ About the storage agents■ DiskGroup agent

■ DiskGroupSnap agent

■ DiskReservation agent

■ Volume agent

■ VolumeSet agent

■ LVMLogicalVolume agent

■ LVMVolumeGroup agent

■ Mount agent

About the storage agentsStorage agents monitor sharedstorage andmake sharedstorage highlyavailable.Storage includes shared disks, disk groups, volumes, and mounts.

DiskGroup agentThe DiskGroup agent brings online, takes offline, and monitors Veritas VolumeManager (VxVM) disk groups. This agent uses VxVM commands. You can use thisagent to monitor or make disk groups highly available.

When thevalueof theStartVolumesandStopVolumes attribute is 1, theDiskGroupagent brings the volumes online and takes them offline during the import anddeport operations of the disk group.

2Chapter



The DiskGroup info agent function gets information from theVolume Manager and displays the type and free size for theDiskGroup resource.

Initiate the info agent function by setting the InfoInterval timingto a value greater than 0.

In the following example, the info agent function executes every60 seconds:

# haconf -makerw

# hatype -modify DiskGroup InfoInterval 60

The command to retrieve information about the DiskType andFreeSize of the DiskGroup resource is:

# hares -value diskgroupres ResourceInfo

Output includes:

DiskType slicedFreeSize 35354136

The value specified is in kilo bytes.

Info

Different action agent functions follow:

■ license.vfdChecks for valid Veritas Volume manager license – if one is notfound use the vxlicinst utility to install a valid license key.

■ disk.vfdChecks if all disks in diskgroup are visible on host — if it fails,check ifthe path todisksexists fromthe host and check ifLUNmasking and zoning are set properly.

■ udid.vfdChecks the UDIDs (unique disk identifiers) of disks on thecluster nodes — if it fails, ensure that the disks that are used forthe disk group are the same on all cluster nodes.

■ verifyplex.vfdChecks if the number of plexes on each site for the CampusCluster setup are set properly — if it fails, check that the sites,disks, and plexes are set properly for a Campus Cluster setup.

■ volinuseChecks if open volumes are in use or file systems on volumesthat are mounted outside of VCS configuration.

See “ High availability fire drill ” on page 30.

Action

Storage agentsDiskGroup agent



Initializes the agent to interface with Intelligent monitoringframework (IMF). The function runs when the agent starts up.

imf_init

Waits fornotificationabout disk groupstate changes. Thefunctionruns after the agent initializes with IMF. The function waits fornotification. Upon receiving notification, the agent takes actionon the resource.

imf_getnotification

Registers the resource entities, which the agent must monitorusing IMF. Thefunction runs for each resource after theresourcegoes into a steady state, either online or offline.

imf_register

State definitions for DiskGroup agent

Indicates that the disk group is imported.ONLINE

Indicates that the disk group is not imported.OFFLINE

Indicates that the disk group has unexpectedly deported or becomedisabled.

FAULTED

Indicates that a problem exists either with the configuration or theability to determine the status of the resource. One cause of this stateis when I/O fencing is not configured — the cluster level attributeUseFence is not set to "SCSI3" but the Reservation attribute value is"SCSI3".

UNKNOWN

Attributes for DiskGroup agentTable 2-1 Required attributes

DescriptionRequiredattribute

Name of the disk group that is configured with Veritas VolumeManager.

Type and dimension: string-scalar

DiskGroup


24



Table 2-2 Optional attributes

DescriptionOptional attributes

If the value is 1, and SCSI-3 fencing is used, the agentmonitors the SCSI reservation on the disk group. If thereservation ismissing, themonitor agent functiontakes the service group containing the resourceoffline.

Type and dimension: boolean-scalar

Default: 0

Note: If the MonitorReservationattribute is set to 0,the value of the clusterwide attribute UseFence is settoSCSI3,andthe disk group is importedwithout SCSI

reservation, then the monitor agent function takesthe service group containing the resource offline.

MonitorReservation




Table 2-2 Optional attributes (continued)


PanicSystemOnDGLoss


26





Determines whether to panic the node if the diskgroup becomes disabled or monitor program timesout. A loss of storage connectivity can cause the diskgroup to become disabled. VxVM commands notresponding properly can cause monitor program totimeout.

Note: System administrators may want to set a highvalue for FaultOnMonitorTimeout to increasesystemtolerance.

This attribute accepts following values 0, 1, 2,3

■ 0 : Do not halt the system■ 1 : halt the system if either disk group goes into

disabled state or the disk group resource faultsdue to monitor timeout

■ 2: halt the system if disk group goes into disabledstate

■ 3: halt thesystemif diskgroup resourcefaultsdueto monitor timeout

If the value of the attribute is 0, and the disk groupbecomes disabled, the following occurs:

■ If the cluster has I/O fencing enabled, theDiskGroup resource is marked FAULTED. Thisstate results in the agent attempting to take theservice group offline.Aspartofbringing theDiskGroupresource offline,the agent attempts to deport the disabled diskgroup. Even if disabled disk group fails to deport,the DiskGroup resource enters a FAULTED state.This stateenables thefailover of theservicegroupthat contains the resource. To fail back theDiskGroup resource, manually deport the diskgroup after restoring storage connectivity.

■ If the cluster does not use I/O fencing, a messageis logged and the resource is reported ONLINE.The resource is reported ONLINE so that it doesnot fail over, which ensures data integrity.

Note: The PanicSystemOnDGLossattribute doesnotdepend on the MonitorReservation attribute.

Note: If PanicSystemOnDGLoss is set to non-zero






value, the system panic is initiated using poweroff-nf command. This command halts the system. Anadministrator needs to bring up the system.

Type and dimension: integer-scalar

Default: 0

If value of this attribute is 1, the DiskGroup onlinefunction starts all volumes belonging to that diskgroup after importing the group.

Note: With VxVM version 5.1.100.0 onwards, if theVeritas Volume Manager defaultautostartvolumesatsystem level is set to on, all the volumes of the diskgroup will be started as a part of the import diskgroup.


Default: 1

StartVolumes

If value is 1, the DiskGroup offline function stops allvolumes belongingto thatdisk group beforeit deportsthe disk group.

Type and dimension: boolean-scalarDefault: 1

StopVolumes

This attribute enables the DiskGroup resource toforcefully gooffline even ifopen volumes aremountedoutside of VCS control. When the value of thisattribute is 1 and the disk group has open volumes,the following occurs:

■ The agent attempts to unmount the file systemson open volumes. If required, the agent attemptsto kill all VCS managed and un-managedapplications using the file systems on those openvolumes.

■ The agent attempts to forcefully unmount thefilesystems to close the volumes.


Default: 0

UmountVolumes


28





Determines if youwant to enable SCSI-3 reservation.This attribute can have one of the following threevalues:

■ ClusterDefault — The disk group is imported withSCSI-3reservation if thevalueof thecluster-levelUseFence attribute is SCSI3. If the value of thecluster-levelUseFenceattribute isNONE,thediskgroup is imported without reservation.

■ SCSI3— The disk group is imported with SCSI-3reservation if the value of the cluster-levelUseFence attribute is SCSI3.

■ NONE— Thedisk group is importedwithoutSCSI-3reservation.


Default: ClusterDefault

Example: "SCSI3"

Reservation

Table 2-3 Internal attributes

DescriptionAttribute

Do not use. For internal use only.tempUseFence

The number of threads that are used within the agent process formanaging resources. This number does not include the number of threads that are used for other internal purposes.

Setting theNumThreads attribute to a higher value may decrease thetime required to go online or the time required to monitor a largenumber of DiskGroup resources.

Type and dimension: static integer-scalar

Default: 1

Note: If there are many DiskGroup resourcesandif theresourcesaretakingmoretime to comeonline, consider increasingtheNumThreadsattribute to a value greater than 1.

NumThreads

The attribute is deprecated.DiskGroupType




Resource type definition for DiskGroup agentThe resource definition for this agent on Linux follows:

type DiskGroup (static keylist SupportedActions = { "license.vfd", "disk.vfd", "udid.vfd","verifyplex.vfd", checkudid, numdisks, campusplex, volinuse,joindg, splitdg, getvxvminfo }static int OnlineRetryLimit = 1static str ArgList[] = { DiskGroup, StartVolumes, StopVolumes, MonitorOnly,MonitorReservation, tempUseFence, PanicSystemOnDGLoss, DiskGroupType,UmountVolumes, Reservation, ConfidenceLevel }static str IMFRegList[] = { DiskGroup, Reservation }static int IMF{} = { Mode = 3, MonitorFreq = 5, RegisterRetryLimit = 3 }str DiskGroupboolean StartVolumes = 1boolean StopVolumes = 1static int NumThreads = 1boolean MonitorReservation = 0temp str tempUseFence = INVALIDint PanicSystemOnDGLoss = 0str DiskGroupType = privateint UmountVolumes = 0str Reservation = ClusterDefault

)

Notes for DiskGroup agentThe DiskGroup agent has the following notes:

■ High availability fire drill

■ Using volume sets

■ Setting the noautoimport flag for a disk group

■ Using the DiskGroup agent with IMF

High availability fire drillThehigh availability firedrill detects discrepancies between theVCSconfigurationand the underlying infrastructure on a node. These discrepancies might preventa service group from going online on a specific node.

For DiskGroup resources, the high availability fire drill checks for:


30



■ The Veritas Volume Manager license

■ Visibility from host for all disks in the disk group

■

The same disks for the disk group on cluster nodes■ Equal numberofplexes on all sites for the diskgroup in a campus cluster setup

For more information about using the high availability fire drill see the VeritasCluster Server Administrator ’ s Guide .

Using volume setsWhen you use a volume set, set StartVolumes and StopVolumes attributes of theDiskGroup resource that contains a volume set to 1. If a file system is created onthe volume set, use a Mount resource to mount the volume set.

See the Mount agent description for more information.

Setting the noautoimport flag for a disk groupVCS requires that the noautoimport flag of an imported disk group be explicitlyset to true. This value enables VCS to control the importation and deportation of disk groups as needed when bringing disk groups online and taking them offline.

To check the status of the noautoimport flag for an imported disk group

# vxprint -l disk_group | grep noautoimport

If the output fromthis command is blank, the noautoimport flag is set to falseand VCS lacks the necessary control.

The Monitor function changes the value of the VxVM noautoimport flag from off to on. It changes the value instead of taking the service group offline. This actionallows VCS to maintain control of importing the disk group.

The following command changes the autoimport flag to false:

# vxdg -g disk_group set autoimport=no

Using the DiskGroup agent with IMFConsiderations to use the DiskGroup agent with IMF:

■ You can either set the MonitorFreq to 0 or a high value. Setting the value of the MonitorFreq key to a high value ensures that the agent does not run themonitor function frequently. Setting the MonitorFreq key to 0 disables the




traditional monitoring while IMF monitoring is in progress. Traditionalmonitoring is done after receiving the notification for a resource.However, if the disk group is configured with reservation and value of the

MonitorReservation attribute is set to 1, then set the MonitorFreq key valueto the frequency at which you want the agent to run the monitor function, toverify the reservation on the disk group.

Sample configurations for DiskGroup agent

DiskGroup resource configurationSample configuration of the DiskGroup resource:

DiskGroup dg1 (DiskGroup = testdg_1

)

Debug log levels for DiskGroup agentThe DiskGroup agent uses the following debug log levels:

DBG_1, DBG_4, DBG_5

DiskGroupSnap agentUse the DiskGroupSnap agent to perform fire drills in a campus cluster. TheDiskGroupSnap agent enables you to verify the configuration and data integrityin a Campus Cluster environment with VxVM stretch mirroring. The agent alsosupports SCSI-3 fencing.

Note: The DiskGroupSnap agent requires the Global Cluster Option (GCO) licenseenabled on all systems in the cluster.

For more information on fire drills, refer to the Veritas Cluster Server

Administrator ’

s Guide .You must define the DiskGroupSnap agent in a separate FireDrill service groupwhich is similar to the Application service group. The FireDrill service groupmight contain resources similar to the Application service group, for exampleMount, Application, and so on.

The FireDrill service group must also contain a resource of type DiskGroupSnapsuch that theMount resource depends on theDiskGroupSnapresource.Themain

Storage agentsDiskGroupSnap agent

32



DiskGroup must contain multiple sites registered in it with the value of the''siteconsistent'' attribute set to on.

When the DiskGroupSnap agent goes online, the agent detaches one of the sitesfrom the main DiskGroup and imports the detached site on the fire drill host asan independent DiskGroup with a different name. The volumes on the DiskGroupare also imported and mounted with same names on the fire drill host.

The DiskGroupSnap agent provides Gold and Bronze configurations for the firedrill,whichcan be specified using theagent'sFDType attribute. Theconfigurationdecides the site to be detached from the DiskGroup for fire drill testing. The Goldconfiguration is the default option in which the agent selects a site from theDiskGroup that is neither the local VxVMsitenor the siteon which the DiskGroupis online. With the Gold configuration, you can also specify an alternate site todetach through the agent's FDSiteName attribute.With the Bronze configuration,

the agent uses the local VxVM site name as the site to detach fromthe DiskGroup.For important information about thisagent, See “ Notes forDiskGroupSnapagent ”

on page 36.

Dependencies for DiskGroupSnap agentTheDiskGroupSnap resourcedoes notdependon anyother resources. Theservicegroup that contains the DiskGroupSnap agent ’ s resource has an offline localdependency on the application ’ s service group. The offline local dependency isto make sure the firedrill service group and the application service group are not

online at the same site at the same time.

Figure 2-2 Sample service group that includes a DiskGroupSnap resource

Offline local dependency

A p p

l i c a

t i o n

s e r v

i c e g r o u p

F i r e

d r

i l l

s e r v

i c e g r o u p Application

Mount

DiskGroupSnap

Application

Mount

DiskGroup




Agent functions for DiskGroupSnap agent

Verifies that the application ’ s disk group is in a valid campus cluster

configuration. It detaches the site that the value of the FDSiteNameattribute specifies. It then creates another disk group to be used forthe fire drill on the detached site. After the completion of Onlinefunction, the agent creates a lock file in the lock directory(/var/VRTSvcs/lock) to indicate that the resource is online.

Online

This re-attaches the site that the value of the FDSiteName attributespecifiesback to theapplication ’ s disk group. After the completion of Offline function theagent removes thelockfile from thelock directory(/var/VRTSvcs/lock) to indicate that the resource is Offline.

Offline

Monitors the DiskGroupSnap resource by checking the existence of

the Lock file in /var/VRTSvcs/lock directory..

Monitor

Takes the DiskGroupSnap resource offline.Clean

If the DiskGroupSnap resource has a parent resource that is notONLINE, then it deletes the online lock file of the DiskGroupSnapresource. This marks the DiskGroupSnap resource as OFFLINE.

Open

State definitions for DiskGroupSnap agent

The DiskGroupSnap resource functions normally.ONLINE

The DiskGroupSnap resource is not running.OFFLINE

A configuration error exists.UNKNOWN

The DiskGroupSnap resource is takenoffline unexpectedly outside of VCS control.

FAULTED

Attributes for DiskGroupSnap agentTable 2-4 Required attributes


The name of the DiskGroup resource from the application servicegroup.

Type-dimension: string-scalar

Example: "dgres"

TargetResName


34



Table 2-4 Required attributes (continued)


Specifies the configuration to be used for the fire drill. The possiblevalues for this attribute are:

■ Bronze■ Gold (default)

TheBronze configuration uses the local host'sVxVM site name as thesite to be detached from the DiskGroup. This action leaves theDiskGroup vulnerable to site disaster since a copy of the productionvolume might not be available when the fire drill is in progress.

In the Gold configuration there are at least three copies of the parent

volume available on different sites, hence, even after detaching onesite the volume is not vulnerable to site disaster while the fire drill isin progress.

FDType


DescriptionOptional attribute

The unique VxVM site name tag for the fire drill disks. Thevalue of this attributeis used inconjunctionwith the FDTypeattribute and it must be set to one of the sites registered inthe main DiskGroup.

■ When FDTypeis set to theBronzeconfiguration, thevalueof FDSiteName should eitherbe empty or thename of thelocal host VxVM site for the fire drill host.

■ When FDType is set to the Gold configuration,FDSiteName identifies a site in the DiskGroup to detachas a part of the fire drill. If FDSiteName is left blank, theagent will choosea site to detach based on theDiskGroupconfiguration. The agent chooses a site name from theDiskGroup which is neither the production server's sitename nor the fire drill host's site name.

Table 2-6 shows the possible values of the attributesFDType and FDSiteName and the decision taken by theagent.

FDSiteName

Consider a configuration where the Production DiskGroup contains three sites:A, B, and C, and the Application service group is online on a node with local VxVMsite ID is A. Fire drill is being done on another node Application service group isonline on a node where local VxVM site ID is B.




Table 2-6 Example FDType configurations

Gold/EmptyBronzeFDType

CBEmptyCBEmptyFDSitename

Detach site Cfrom theDiskGroup

ErrorCheck if there isanother siteother thanA and B andselect it.Else, it is anerror

ErrorDetach site B fromDiskGroup

Use B asthesite todetach andproceed

Result

Notes for DiskGroupSnap agentThe DiskGroupSnap agent has the following notes:

■ See “ Fire drill configuration after upgrading VCS ” on page 36.

■ See “ Configuring the SystemZones attribute for the fire drill service group ”

on page 36.

■ See “ Configuring the FireDrill service group ” on page 37.

■ See “ Adding the ReuseMntPt attribute to the ArgList attribute for the Mountagent type ” on page 37.

■ See “ Configuration considerations ” on page 38.

■ See “ Agent limitations ” on page 39.

Fire drill configuration after upgrading VCSAfter upgrading VCS from any earlier version to 6.0, delete all resources of typeDiskGroupSnapandrecreate them again using thenewdefinitionsof theattributes.Failure to perform this step might result in an unexpected behavior of the agent.

Configuring the SystemZones attribute for the fire drill servicegroupYou must assign the local system values to the SystemZones attribute of theapplication ’ s service group. You set these values so that the service group failsover in the same zone before it tries to fail over across zones.

For more information about campus cluster setup, refer to the Veritas Cluster Server Administrator ’ s Guide .


36



For example, you set up the service group ’ s SystemZones attribute for two zones:0 and 1. You want the service group on Node_A and Node_B to fail over betweenthe two nodes before it comes up on Node_C and Node_D. The application and its

fire drill service group both have the following values for the SystemZonesattribute:

SystemZones = { Node_A = 0, Node_B = 0, Node_C = 1, Node_D = 1 }

Configuring the FireDrill service groupIn theFireDrill service group, theapplication-level resources (for example, processresources, application resources, or Oracle resources, and so on) can have thesame attribute values in the firedrill service group and the application servicegroup. The reuse of the same values for the attributes can result in VCS reporting

the wrong resources as online.Set theFireDrill type-level attribute to 1 for those types. For example, if theOracleand Listener resources are configured identically, set the FireDrill attribute forOracle and Netlsnr to 1:

# haconf –makerw# hatype –modify Oracle FireDrill 1# hatype –modify Netlsnr FireDrill 1# haconf -dump –makero

Adding the ReuseMntPt attribute to the ArgList attribute forthe Mount agent typeIf you plan to use a Mount resource in a firedrill service group, you must add theReuseMntPt attribute to ArgList and set its value to 1.

To add the ReuseMntPt attribute to the ArgList attribute and set its value to 1

1 Make the configuration read and write.

# haconf -makerw

2 Add the ReuseMntPt attribute to the ArgList attribute.

# hatype -modify Mount ArgList -add ReuseMntPt

3 Change the value of the ReuseMntPt attribute to 1 for the firedrill ’ s Mountresource.

# hares -modify firedrill_mount_resource_name ReuseMntPt 1




4 Change the value of the ReuseMntPt attribute to 1 for the original Mountresource.

# hares -modify original_mount_resource_name ReuseMntPt 1

5 Make the configuration read only.

# haconf –dump –makero

Configuration considerationsKeep the following recommendations in mind:

■ You must install Veritas Volume Manager 5.1 or later with the FMR licenseand the Site Awareness license.

■ Do not bring the DiskGroupSnap resource online in the SystemZone wherethe application service group is online.

■ Make sure that the firedrill service group and the application service groupboth use the same values for the SystemZones attribute.

■ Donot use Volume resources in thefiredrill servicegroup. TheDiskGroupSnapagent internallyuses the vxvol command to start all the volumes in the firedrilldisk group.

■

In large setups, you may need to tweak the various timer values so that thetimers do not time out while waiting for VxVM commands to complete. Thetimers you need to tweak are the OfflineTimeout for the DiskGroupSnapresourceandMonitorInterval andActionTimeout for theassociatedDiskGroupresource, for example:

# haconf –makerw# hares -override dgsres OfflineTimeout# hares -modify dgsres OfflineTimeout 600# hares -override dgres MonitorInterval# hares -modify dgres MonitorInterval 1200 (this has to be twice

the value intended for ActionTimeout below)# hares -override dgres ActionTimeout# hares -modify dgres ActionTimeout 600# haconf –dump –makero

■ When you create the firedrill service group, in general use the same attributevalues that you use in the application service group.


38



The BlockDevice attribute of the Mount resource changes between theapplication service group and the firedrill service group. In the BlockDevicepath, you must append an _fd to the disk group name portion, for example,

/dev/vx/dsk/newdg1/newvol1 becomes /dev/vx/dsk/newdg1_fd/newvol1.See Figure2-3 on page39. shows thechanges to resourcevalues for thefiredrillservice group; note that the Volume resource is not included.

■ Before commencing the fire drill, make sure that all the sites registered in theapplication DiskGroup are in ACTIVE state.

Figure 2-3 Sample resource values for a DiskGroupSnap resource

procfdres1

BlockDevice =" /dev/vx/dsk/newdg1_fd/newvol1"

mntfdres1

dgsres

Agent limitationsThe following limitations apply to the DiskGroupSnap agent:

■ The DiskGroupSnap agent does not support Volume Sets.

■ The DiskGroupSnap agent cannot be used in a Storage Foundation RACenvironment.

■ Theonlineandoffline operations of theDiskGroupSnap resource invokes VCSaction entry points to run VxVM commands to detach/reattach the fire drillsite. Since VxVM requires that these commands are run on the node wherethe disk group is imported, the disk group has to be imported on some nodein the cluster before these operations.

■ Take the firedrill service group offline before you shut down VCS on any node.If you fail to take the firedrill service group offline before you shut down VCS,you must manually reattach the fire drill site to the disk group to continue toperform fire drills.

■ Use the enclosures that have the ASL/APM libraries that are supported in theVeritas VolumeManager. To view thesupportedenclosures, use the vxddladm

listsupport command.

■ Do not switch the Application service group when fire drill is in progress.




Resource type definition for DiskGroupSnap agentThe resource type definition for this agent follows:

type DiskGroupSnap (static int ActionTimeout = 120static int MonitorInterval = 300static int NumThreads = 1static str ArgList[] = { TargetResName, FDSiteName, FDType }str TargetResNamestr FDSiteNamestr FDType)

Sample configurations for DiskGroupSnap agentIn Figure 2-4 , the Primary site is in the Bronze configuration and the Disasterrecovery site is in a Gold configuration.

Since the Primary site does not have dedicated fire drill disks, it is in a Bronzeconfiguration. In the Bronze configuration, you re-purpose the mirror disks inthe disaster recovery site to serve as fire drill test disks. The drawback with theBronze configuration is that if a disk failure occurs when the fire drill is onlineat the Primary site, it results in a site failure.

The FDSiteName value in a bronze configuration is the VxVM site name. For this

configuration, the FDSiteName attribute values for the nodes at the Primary sitefollow:

FDSiteName@Node_A = priFDSiteName@Node_B = pri

The Disaster Recovery site is in a Gold configuration as it has dedicated fire drilldisks at the site. For the FDSiteName attribute, use the VxVM site tag given tothe fire drill disks. For this configuration, the FDSiteName attribute values forthe nodes at the Disaster recovery site follow:

FDSiteName@Node_C = dr_fdFDSiteName@Node_D = dr_fd

Set values for the SystemZones attribute to zero for Node_A and Node_B, and onefor Node_C and Node_D. For example:

SystemZones = { Node_A = 0, Node_B = 0, Node_C = 1, Node_D = 1 }


40



Figure 2-4 Primarysitewith theBronze configuration and thedisaster recoverysite with the Gold configuration

Primary site in a Bronze

configuration

Disaster recovery site in

a Gold configuration

Stretchcluster

Node_DNode_CNode_BNode_A

VxVM sitename = pri

SystemZones=0 SystemZones=1

Data disks Data disks

Fire drill disks

VxVM sitename=pri VxVM sitename=dr

VxVM sitename = dr

VxVM sitename = dr_fd

Typical main.cf configuration for DiskGroupSnap agentThe following sample configuration shows the fire drill ’ s service group and itscorresponding application service group. The fire drill ’ s service group follows:

group dgfdsg (SystemList = { Node_A = 0, Node_B = 1, Node_C = 2, Node_D = 3 }SystemZones = { Node_A = 0, Node_B = 0, Node_C = 1, Node_D = 1})




DiskGroupSnap dgsres (TargetResName = dgres

)

FDtype = "Gold"FDSiteName @Node_A = priFDSiteName @Node_B = priFDSiteName @Node_C = dr_fdFDSiteName @Node_D = dr_fd)

Mount mntfdres1 (MountPoint = "/dgsfs1"

BlockDevice = "/dev/vx/dsk/newdg1_fd/newvol1"FSType = vxfsFsckOpt = "-y"ReuseMntPt = 1)

Mount mntfdres2 (MountPoint = "/dgsfs2"BlockDevice = "/dev/vx/dsk/newdg1_fd/newvol2"FSType = vxfsFsckOpt = "-y"ReuseMntPt = 1)

Process procfdres1 (PathName = "/usr/bin/ksh"Arguments = "/scrib.sh /dgsfs1")

Process procfdres2 (PathName = "/usr/bin/ksh"

Arguments = "/scrib.sh /dgsfs2")

requires group dgsg offline local

mntfdres1 requires dgsresmntfdres2 requires dgsres


42



procfdres1 requires mntfdres1

procfdres2 requires mntfdres2

The application ’ s service group (the actual service group) follows:

group dgsg (SystemList = { Node_A = 0, Node_B = 1, Node_C = 2, Node_D = 3 }SystemZones = { Node_A = 0, Node_B = 0, Node_C = 1, Node_D = 1})

DiskGroup dgres (DiskGroup = newdg1)

Mount mntres1 (MountPoint = "/dgsfs1"BlockDevice = "/dev/vx/dsk/newdg1/newvol1"FSType = vxfsFsckOpt = "-y"ReuseMntPt = 1)

Mount mntres2 (

MountPoint = "/dgsfs2"BlockDevice = "/dev/vx/dsk/newdg1/newvol2"FSType = vxfsFsckOpt = "-y"

ReuseMntPt = 1)

Process procres1 (PathName = "/usr/bin/ksh"Arguments = "/scrib.sh /dgsfs1")

Process procres2 (PathName = "/usr/bin/ksh"Arguments = "/scrib.sh /dgsfs2")

mntres1 requires dgresmntres2 requires dgres




procres1 requires mntres1procres2 requires mntres2

Sample main.cf of DiskGroupSnap with Oracle resourceThe following Oracle configuration has been simplified for presentation withinthis guide.

group fd_oragrp (SystemList = { Node_A = 0, Node_B = 1 }AutoStart = 0SystemZones = { Node_A = 0, Node_B = 1 })

DiskGroupSnap dgres (FDSiteName @Node_A = siteAFDSiteName @Node_B = siteBTargetResName = oradg_resFDType = "Bronze")

IP fd_oraip (Device = ethoAddress = "10.198.95.191"NetMask = "255.255.255.0"

)

Mount fd_archmnt (FsckOpt = "-y"ReuseMntPt = 1BlockDevice = "/dev/vx/dsk/oradg_fd/archive_vol"MountPoint = "/ora_archive"FSType = vxfs)

Mount fd_datamnt (FsckOpt = "-y"ReuseMntPt = 1BlockDevice = "/dev/vx/dsk/oradg_fd/data_vol"MountPoint = "/ora_data"FSType = vxfs)

NIC fd_oranic (


44



Device = eth0NetworkHosts = { "10.198.95.1" })

Netlsnr fd_LSNR (Home = "/opt/oracle/ora_home"Owner = oracle)

Oracle fd_Ora_01 (Owner = oracleHome = "/opt/oracle/ora_home"Sid = Ora_01)

requires group oragrp offline localfd_LSNR requires fd_Ora_01fd_LSNR requires fd_oraipfd_Ora_01 requires fd_archmntfd_Ora_01 requires fd_datamntfd_archmnt requires dgresfd_datamnt requires dgresfd_oraip requires fd_oranic

group oragrp (

SystemList = { Node_A = 0, Node_B = 1 }AutoStartList = { Node_A, Node_B }SystemZones = { Node_A = 0, Node_B = 1 })

DiskGroup oradg_res (DiskGroup = oradg)

IP Node_A4vip (Device = eth0Address = "10.198.95.192"Netmask = "255.255.252.0"

)

Mount arch_mnt (FsckOpt = "-y"ReuseMntPt = 1BlockDevice = "/dev/vx/dsk/oradg/archive_vol"




MountPoint = "/ora_archive"FSType = vxfs)

Mount data_mnt (FsckOpt = "-y"ReuseMntPt = 1BlockDevice = "/dev/vx/dsk/oradg/data_vol"MountPoint = "/ora_data"FSType = vxfs)

NIC nic_Node_A4vip (Device = eth0

)

Netlsnr LSNR (Home = "/opt/oracle/ora_home"Owner = oracle)

Oracle Ora_01 (Owner = oracleHome = "/opt/oracle/ora_home"

Sid = Ora_01)

Volume arch_vol (Volume = archive_volDiskGroup = oradg)

Volume data_vol (Volume = data_volDiskGroup = oradg)

LSNR requires Ora_01LSNR requires Node_A4vipOra_01 requires arch_mntOra_01 requires data_mntarch_mnt requires arch_volarch_vol requires oradg_resdata_mnt requires data_vol


46



Indicates that the number of reserved disks is greater than or equalto the percentage specified in the resource definition.

ONLINE

Disks are not reserved.OFFLINE

Indicates that a problem exists with the configuration.UNKNOWN

Attributes for DiskReservation agentTable 2-7 Required attributes


A list of raw disk devices. Use the absolute or relative device path.

The absolute or relative device path allows a maximum of 64characters. The relative path is assumed to start from the /devdirectory.

The order of the disks in the list must be the same across all systemsin the cluster, evenif thesamedevicehas a different nameondifferentsystems.

Note: You must change this attribute before bringing a resourceonline. An online device must be taken offline before altering thisattribute because disk reservation occurs during the process of bringing a resource online.

Type and dimension: string-vector

Example: "/dev/sdd"

Disks


DescriptionOptionalattribute

If enabled, FailFast causes the system to panic when a reservationconflict is detected, reducing the chance of further data corruption.


Default: 0

FailFast

Storage agentsDiskReservation agent

48





Minimum percentage of configured disks that can be reserved beforea resource cango online. The percentagemustbe greater than or equalto 51, and less than or equal to 100.

If the value specified is less than 51, the percentage is set to 51.

If the value specified is greater than 100, the percentage is set to 100.


Default: 100

Percentage

Alters the periodicity (in seconds) of the automatic probe function

that checks the reservation status of the disks. The value must begreater than or equal to three, and less than or equal to 15.

If the value specified is less than 3, the interval is set to 3.

If the value specified is greater than 15, the interval is set to 15.

A lower value for ProbeInterval specifies more frequent probes andprovides for quicker discovery of reservation conflicts. SymantecCorporation recommends a value is between 3 and 8.


Default: 3

ProbeInterval

Resource type definition for DiskReservation agentThe resource definition for this agent on Linux follows:

type DiskReservation (static str ArgList[] = { Disks, FailFast, Percentage,ProbeInterval }str Disks[]boolean FailFast = 0int Percentage = 100int ProbeInterval = 3

)

Notes for DiskReservation agentThe DiskReservation agent has the following notes:

■ The DiskReservation agent does not reserve disks that have multiple paths




■ Configuring the MonitorTimeout attribute for more than three disks

The DiskReservation agent does not reserve disks that havemultiple pathsYou cannot use the DiskReservation agent to reserve disks that have multiplepaths. In case of Veritas Dynamic Multi-Pathing, the LVMVolumeGroup and theLVMLogicalVolume agents can be used without the DiskReservation agent. Toensure data protection on such a configuration, Symantec recommends the useof Veritas VolumeManager (VxVM)disk groups. Note that VxVM requires theuseof SCSI-3 compliant disks.

Configuring theMonitorTimeout attribute for more than threedisksThe MonitorTimeout attribute ’ s setting of 60 is adequate for up to three disks.When configuring the MonitorTimeout attribute for more than three disks, usethe following formula:

Set MonitorTimeout to be equal orgreater than 15 times thetotal numberof disks.(MonitorTimeout >= 15 * Number of disks).

For example, if you have eight disks, MonitorTimeout is 120 or greater.

Sample configurations for DiskReservation agentThe sample configurations for this agent agent follow:

Configuration 1 for DiskReservation agentIn this example, the DiskReservation agent reserves a disk. The disk is mountedwith the Veritas File System.

system sysA

system sysB

group groupx (SystemList = { sysA, sysB }AutoStartList = { sysA })

DiskReservation diskres1 (Disks = { "/dev/sdc" }FailFast = 1


50



)

Mount mount (

MountPoint = "/mnt/tmp"BlockDevice = "/dev/sdc"FSType = vxfsMountOpt = rw)

mount requires diskres1

// resource dependency tree//// group groupx

// {// Mount mount// {// DiskReservation diskres1// }// }

Debug log levels for DiskReservation agentThe DiskReservation agent uses the following debug log levels:

DBG_1, DBG_3, DBG_4, DBG_5.

Volume agentThe Volume agent brings online, takes offline, and monitors a Veritas VolumeManager (VxVM) volume. Use the agent to make a volume highly available.

Note: Do not use the Volume agent for volumes created for replication.

Dependencies for Volume agentVolume resources depend on DiskGroup resources.

Storage agentsVolume agent



Figure 2-5 Sample service group that includes a Volume resource

Application

IP Mount

NIC Volume

DiskGroup

Agent functions for Volume agentUses the vxrecover command to start the volume.Online

Uses the vxvol command to stop the volume.Offline

Attempts to read a block from the raw device interface to the volumeto determine if the volume is online, offline, or unknown.

Monitor

Terminates all ongoing resource actions and takes the resourceoffline — forcibly when necessary.

Clean

State definitions for Volume agent

Indicates that thespecified volumeis startedand that I/Ois permitted.ONLINE

Indicates that the specified volume is not started and that I/O is notpermitted.

OFFLINE

Indicates the volume stopped unexpectedly and that I/O is notpermitted.

FAULTED

Indicates that the agent could notdetermine thestate of theresource

or that the resource attributes are configured incorrectly.

UNKNOWN


52



Attributes for Volume agentTable 2-9 Required attributes


Name of the disk group that contains the volume.


Example: "DG1"

DiskGroup

Name of thevolume from disk group specifiedin DiskGroupattribute.


Example: "DG1Vol1"

Volume

Table 2-10 Internal attribute


Number of threads used within the agent process for managingresources. This number does not include threads used for otherinternal purposes.

Do not modify this attribute for this agent.

Setting this attribute to a higher value may result in agent functiontimeouts due to serialization of underlying commands.

Default: 1

NumThreads

Resource type definition for Volume agentThe resource type definition for this agent follows:

type Volume (static int NumThreads = 1static str ArgList[] = { Volume, DiskGroup }str Volumestr DiskGroup

)

Sample configuration for Volume agentThe sample configuration for the Volume agent follows:




Volume sharedg_vol3 (Volume = vol3DiskGroup = sharedg

)

Debug log levels for Volume agentThe Volume agent uses the following debug log levels:

DBG_1, DBG_3, DBG_5

VolumeSet agentThe VolumeSet agent brings online, takes offline, and monitors a Veritas VolumeManager (VxVM)volume set. Use theagent to make a volumeset highlyavailable.

Dependencies for VolumeSet agentVolumeSet resources depend on DiskGroup resources.

Figure 2-6 Sample service group that includes a VolumeSet resource

Application

IP Mount

NIC VolumeSet

DiskGroup

Agent functions for VolumeSet agent

Uses the vxrecover command to start the volume set.Online

Uses the vxvset command to stop the volume set.Offline

Attempts to read a block from therawdeviceinterface to thevolumesinside the volume set to determine if the volume set is online, offline,or unknown.

Monitor

Storage agentsVolumeSet agent

54



Terminates all ongoing resource actions and takes the resourceoffline — forcibly when necessary.

Clean

State definitions for VolumeSet agent

Indicates that all the volumes in the volume set are started and thatI/O is permitted for all the volumes.

ONLINE

Indicates that at least one of the volume is not started in the volumeset and that I/O is not permitted for that volume.

OFFLINE

Indicates the volumes that are inside the volume set have stoppedunexpectedly and that I/O is not permitted.

FAULTED

Indicates that the agent could notdetermine thestate of theresourceor that the resource attributes are configured incorrectly.

UNKNOWN

Attributes for VolumeSet agentTable 2-11 Required attributes


The name of the disk group that contains the volume set.


Example: "DG1"

DiskGroup

The name of the volume set from the disk group that you specified inthe DiskGroup attribute.


Example: "DG1VolSet1"

VolumeSet

Resource type definition for VolumeSet agent

type VolumeSet (static str ArgList[] = { DiskGroup, VolumeSet }str VolumeSetstr DiskGroup

)

Storage agentsVolumeSet agent



Sample configurations for VolumeSet agentThis sections contains sample configurations for this agent.

A configured VolumeSet that is dependent on a DiskGroupresourceThe VolumeSet ’ s shared_vset3 resource is configured and is dependent onDiskGroup resource with a shared diskgroup.

VolumeSet sharedg_vset3 (VolumeSet = vset3DiskGroup = sharedg)

Agent notes for VolumeSet agentThis sections contains notes about this agent.

Inaccessible volumesprevent the VolumeSet agent from coming onlineThe VolumeSet agent does not come online if any volume is inaccessible in itsvolume set.

To remove a volume from volume set

Enter thefollowingcommands to removea volumefrom a volumesetmountedon mountpoint.

# fsvoladm remove mountpoint volume_name# vxvset -g diskgroup rmvol volumeset volume_name

Debug log levels for VolumeSet agentThe VolumeSet agent uses the following debug log levels:

DBG_1, DBG_4

LVMLogicalVolume agentTheLVMLogicalVolume agent brings online, takes offline, andmonitors a LogicalVolume Manager (LVM2) volume. This agent uses LVM2 commands. You can usethis agent to make volume groups and logical volumes highly available and tomonitor them.

Storage agentsLVMLogicalVolume agent

56



DependenciesLVMLogicalVolume resources depend on LVMVolumeGroup resources.

Figure 2-7 Sample service group that includes a LVMLogicalVolume resource

Application

IP Mount

NIC LVMLogicalVolume

LVMVolumeGroup

Agent functions

Starts the volume using the lvchange command.Online

Stops the volume using the lvchange command.Offline

Determines if the volume is online or offline by reading a block from

the raw device interface to the volume.

Monitor

Terminates allongoing resource actionsand takes theresource offline,forcibly when necessary.

Clean

State definitionsThe state definitions for this agent follow:

Indicates that thespecifiedvolumeis startedandthat I/Ois permitted.ONLINE

Indicates that the specified volume is not started—

and I/O is notpermitted.OFFLINE

Indicates that the agent could notdetermine thestate of theresourceor that the resource attributes are invalid.

UNKNOWN




AttributesTable 2-12 Required attributes


Name of the volume that is configured with Logical Volume Manager(LVM2).


Example: "volume1"

LogicalVolume

Name of the volume group that is configured with Logical VolumeManager (LVM2), which contains the volume.

Type and dimension: string-scalarExample: "volumegroup1"

VolumeGroup

Resource type definitionThe resource definition for this agent on Linux follows:

type LVMLogicalVolume (static str ArgList[] = { LogicalVolume, VolumeGroup }str LogicalVolume

str VolumeGroup)

Sample configurationsThe sample configuration for the LVMLogicalVolume agent follows:

Configuration 1In this example, /dev/sdc and /dev/sdd are the disks where the volume grouptestvg_1 is created.

LVMLogicalVolume lvol1 (LogicalVolume = testvol_1VolumeGroup = testvg_1

)LVMVolumeGroup lvg1 (

VolumeGroup = testvg_1)DiskReservation dr1 (


58



Disks = { "/dev/sdc", "/dev/sdd" })lvol1 requires lvg1

lvg1 requires dr1

Configuration 2In this example, you use the DiskReservation resource to verify that disks areavailable only to one system. The LVM2 logical volumes on the LVM2 volumegroups that are imported are started if the reservation is confirmed. The logicalvolumes can then be mounted at a mount point.

DiskReservation dr_cde (Disks = { "/dev/sdc", "/dev/sdd", "/dev/sde" }

)Mount mnt_lvmvol01 (MountPoint = "/mnt/lvmvol01"BlockDevice = "/dev/mapper/lvmvg01-lvmvol01"FSType = "reiserfs"FsckOpt = "-y")

LVMLogicalVolume lvmvol01 (LogicalVolume = lvmvol01VolumeGroup = lvmvg01)

LVMVolumeGroup lvmvg01 (VolumeGroup = lvmvg01)

mnt_lvmvol01 requires lvmvol01lvmvol01 requires lvmvg01

lvmvg01 requires dr_cde

Debug log levelsThe LVMLogicalVolume agent uses the following debug log levels:

DBG_1, DBG_3, DBG_5

LVMVolumeGroup agentThe LVMVolumeGroup agent brings online, takes offline, and monitors a LogicalVolume Manager (LVM2) volume group. This agent uses LVM2 commands. Youcan use this agent to make volume groups and logical volumes highly availableand to monitor them.

Storage agentsLVMVolumeGroup agent



This agent supports Veritas Dynamic Multi-Pathing.

Dependencies for LVMVolumeGroup agentNo fixed dependencies exist for LVMVolumeGroup Agent. When you create avolume group on disks with single path, Symantec recommends that you use theDiskReservation agent.

With Veritas Dynamic Multi-Pathing, do not use the DiskReservation agent.

Figure 2-8 Sample service group that includes a LVMVolumeGroup resource

Application

IP Mount

NIC LVMLogicalVolume

LVMVolumeGroup

Agent functions for LVMVolumeGroup agentThe agent functions for this agent follow:

Imports the volume group using the vgimport command.Online

Exports the volume group using the vgexport command.Offline

Determines if the volume group is online or offline using thevgdisplay command.

Monitor


Clean


■ volinuseChecks if the specified volume is mounted outside of VCS controlor not. In case it is mounted outside of VCS control, the actionreturns "FAILURE", otherwise it returns "SUCCESS".

Action


60



State definitions for LVMVolumeGroup agent

Indicates that the volume group is imported.ONLINE

Linux: Indicates that the volume group is not imported.OFFLINE

Indicates that a problem exists either with the configuration or theability to determine the status of the resource.

UNKNOWN

Attributes for LVMVolumeGroup agentTable 2-13 Required attributes

DescriptionRequired

attributeThename of thevolume group that is configured with Logical VolumeManager (LVM2) that contains the volume.


Example: "volumegroup1"

VolumeGroup



If thevalueof this attribute is1, theLVMVolumeGroup onlinefunctionimports the group. It then starts all the volumes that belong to thatvolume group.


Default: 0

StartVolumes

If thevalueof this attribute is1, theLVMVolumeGroup onlinefunctionassociates an LVM tag with the volume group and then imports thegroup. While taking the resource offline, it removes the associatedtags.


Default: 0

EnableLVMTagging

Resource type definition for LVMVolumeGroup agentThe resource definition for this agent on Linux follows:




type LVMVolumeGroup (static keylist SupportedActions = { volinuse }static str ArgList[] = { VolumeGroup, StartVolumes,

EnableLVMTagging }str VolumeGroupboolean StartVolumes = 0boolean EnableLVMTagging = 0)

LVMVolumeGroup agent notesThe LVMVolumeGroup agent has the following notes:

Enabling volume group activation protection for Veritas DynamicMulti-PathingOn each node in thecluster, perform the following procedure to enableactivationprotection for volume groups on Red Hat and SUSE systems.

To enable volume group activation protection

1 On each node in the cluster, edit /etc/lvm/lvm.conf , and add the followingline:

tags { hosttags = 1 }

2 On each node in the cluster, create the file lvm_`uname – n`.conf in the /etc/lvm/ directory.

3 Add the following line to the file you created in step 2:

activation { volume_list="@node" }where node is the value of the uname -n command.

Notes for volume group activation protection

Review the following notes if you plan to usevolume group activation protection:■ Setting hosttag=0 in /etc/lvm/lvm.conf disables the lvm tagging for all the

volume groups.

■ If the system ’ s LVM configuration (/etc/lvm/lvm.conf) is modified to supportLVM tagging (hosttags = 1 in the lvm.conf file), make sure that the value of EnableLVMTagging is 1 for all the configured LVMVolumeGroup resources.


62



Another way to disable the LVM tagging for a particular volume group is toadd that volume group name in a volume_list activation configuration ( /etc/lvm/lvm_`uname – n`.conf).

For example, if you donot want LVM taggingto beenforcedfor thevg11volumegroup , then add the following line in /etc/lvm/lvm_`uname – n`.conf. .activation{ volume_list=["vg11","@node"]} where "node"is thevalueof "uname-n" commandUsingthis configurationinstructs theagent to not use tagging whileactivatingthevolumegroup vg11. Therest of thevolumegroupsrequire tagging however.

Sample configurations for LVMVolumeGroup agentThe sample configurations for this agent follows:

Linux configuration 1In this example, /dev/sdc and /dev/sdd are the disks with a single path where thevolume group testvg_1 is created.

LVMVolumeGroup lvg1 (VolumeGroup = testvg_1

)

DiskReservation dr1 (Disks = { "/dev/sdc", "/dev/sdd" }

)

lvg1 requires dr1

Linux configuration 2In this example, the DiskReservation resource is used to verify that disks areavailable only to one system. All LVM2 logical volumes on the LVM2 volumegroups that are imported are started if the reservation is confirmed. You can thenmount the logical volumes at a mount point.

DiskReservation dr_cde (Disks = { "/dev/sdc", "/dev/sdd", "/dev/sde" })

Mount mnt_lvmvol01 (MountPoint = "/mnt/lvmvol01"BlockDevice = "/dev/mapper/lvmvg01-lvmvol01"FSType = "reiserfs"




FsckOpt = "-y")

LVMVolumeGroup lvmvg01 (VolumeGroup = lvmvg01StartVolumes = 1)

mnt_lvmvol01 requires lvmvg01lvmvg01 requires dr_cde

Linux configuration 3In this example, the volume group testvg_1 is created on disks that have multiple

paths and have LVM tagging enabled.

LVMVolumeGroup lvg1 (VolumeGroup = testvg_1EnableLVMTagging = 1)

Linux Configuration 4In this example, the volume groups testvg_1 and testvg_2 are created on disksthat have multiple paths. LVM tagging is enabled for testvg_1 and disabled for

test_vg2.

LVMVolumeGroup lvg1 (VolumeGroup = testvg_1EnableLVMTagging = 1)LVMVolumeGroup lvg2 (VolumeGroup = testvg_2)Here the /etc/lvm/lvm_`uname –n .̀conf file contains:activation { volume_list=["testvg_2","@node"] }

where node is the value of uname -n command.

Debug log levels for LVMVolumeGroup agentThe LVMVolumeGroup agent uses the following debug log levels:

DBG_1, DBG_3, DBG_5


64



Mount agentThe Mount agent brings online, takes offline, and monitors a file system or an

NFS client mount point. You can use the agent to make file systems or NFS clientmount points highly available. This agent is IMF-aware and uses asynchronousmonitoring framework (AMF) kernel driver for IMF notification. For moreinformation about the Intelligent Monitoring Framework (IMF) and intelligentresource monitoring, refer to the Veritas Cluster Server Administrator ’ s Guide .

Note: Intelligent Monitoring Framework for mounts is supported only for thefollowing mount types: VxFS and NFS.

This agent also supports high availability fire drills.

For important information about this agent, See “ Notes for Mount agent ”

on page 73.

Dependencies for Mount agentThe Mount resource does not depend on any other resources.

Figure 2-9 Sample service group that includes a Mount resource

Application

IP Mount

NIC Volume

DiskGroup

Agent functions for Mount agent

Mounts a block device on the directory. If the mount process fails fornon-NFS mounts, theagent attempts to run thefsck command on thedevice before attempting to mount the file system again.

If file system type is NFS, agent mounts the remote file system to aspecified directory. The remote NFS file system is specified in theBlockDevice attribute.

Online

Storage agentsMount agent



Unmounts the mounted file system.Offline

Determines if the file system is mounted.

If IMF is enabled for the Mount agent, the resource is monitoredasynchronously and any change in the resource state is immediatelysent to VCS for appropriate action.

Monitor

Initializes the agent to interface with the asynchronous monitoringframework (AMF) kernel driver. This function runs when the agentstarts up.

imf_init

Waits for notification about resource state changes. This functionruns after the agent initializes with theAMF kernel driver. Theagentcontinuously waits for notification and takes action on the resourceupon notification.

imf_getnotification

Registers the resource entities, which the agent must monitor, withthe AMF kernel driver. This function runs for each resource after theresource goes into steady state (online or offline). This action entrypoint registers themount point,block device, andfile systemtype formount agent.

imf_register

Unmounts the mounted file system forcefully.Clean

The Mount agent info function executes the command:

# d f -h mount_point

The output displays Mount resource information:Size Used Avail Use%

To initiate the info agent function, set the InfoInterval timing to avaluegreater than 0. In this example, theinfoagentfunctionexecutesevery 60 seconds:

# haconf -makerw

# hatype -modify Mount InfoInterval 60

The command to retrieve information about the Mount resource is:

# hares -value mountres ResourceInfoOutput includes:

Size 2097152Used 139484Available 1835332Used% 8%

Info


66



Attributes for Mount agentTable 2-15 Required attributes


Block device for mount point.

For LVM2, use the actual mapper path to the volume.


Examples:

■ LVM2 example

# ls -la /dev/ora_vg/ora_vollrwxrwxrwx 1 root root 26 Apr 17 04:48/dev/ora_vg/ora_vol -> /dev/mapper/ora_vg-ora_vol

Use the path /dev/mapper/ora_vg-ora_vol for the BlockDeviceattribute.

■ NFS exampleIf the file system type is NFS, then specify the BlockDevice as:server:/path/to/sharegalaxy.symantecexample.com:/usr/share1

■ "/dev/vx/dsk/myvcs_dg/myvol"

BlockDevice

Mandatory for non-NFS mounts.

Usethis attributeto specify options forthe fsck command. Youmustcorrectly set this attributefor local mounts. If themount process fails,the fsck command is executed with the specified options before itattempts to remount the block device. Its value must include either-y or -n . Refer to the fsck manual page for more information.

The -y argument enablesthe VxFSfilesystems to performa log replaybefore a full fsck operation.

For NFS mounts, the value of this attribute is not applicable and is

ignored.Type and dimension: string-scalar

VxFS example: -y

Note: When you use the command line, add the % sign to escape '-'.For example: hares -modify MntRes FsckOpt %-y

FsckOpt


68





Type of file system.

Supports vxfs, bind, ext2, ext3, ext4, xfs, nfs, or reiserfs.


FSType

Directory for mount point.


Example: "/mnt/apache1"

MountPoint



If thevalue of this attribute is1, this attribute automatically unmountsVxFS Storage Checkpoints when the file system is unmounted.

If thevalueof this attribute is0, andStorageCheckpointsaremounted,then failover does not occur.


Default: 1

CkptUmount

Options for the mount command. Refer to the mount manualpage formore information.

Do not specify -o in the MountOpt field.

The agent uses this option only when bringing a Mount resourceonline.

If the value of this attribute is nfs, the agent always mounts the filesystem with the "soft,intr" option.


Example: "rw"

MountOpt






This attribute is applicable to Veritas (VxFS) file systems. It controlsthe agent's use of the locking feature provided by vxfs to preventaccident unmounts.

If the value of this attribute is 0, the agent does not lock the mountpoint when the resource is brought online. It does not monitor thestatus of the lock when the resource is online. No warnings appear if the mount has been locked with a key different than "VCS".

If the value of this attribute is 1, during online, the agent uses the key"VCS" to lock the mount point. The monitor agent function monitorsthe locks during every cycle.

■ If the mountpoint is not locked, the agent locks it.■ If the mountpoint is already locked with a key other than "VCS",

the agent logs a warning. It then requests that you run theChgmntlock action agent function.

During offline, the agent, as required, unlocks using whatever keyneeded.


Default: 0

VxFSMountLock

If thevalue of this attribute is1, this attribute automatically unmountsVxFS snapshots when the file system is unmounted.

If the value of this attribute is 0, and snapshots are mounted, theresource cannot be brought offline. In this case, failover does notoccur.


Default: 0

SnapUmount

If the value of this attribute is set to 1 or 2, the monitor verifies thatthe values of the MntPtPermission, MntPtOwner, and MntPtGroupattributes are the same as the actual mounted file system values. If any of these donot match the valuesthatyou havedefined, a messageis logged.

If the value of this attribute is 2, and if the mounted file systempermissions do not match the attribute values, the Monitor agentfunction returns the state as OFFLINE.


Default: 0

AccessPermissionChk


70





If the value ofthisattributeis 0,nomountpoint iscreated.The mountcan fail if the mount point does not exist with suitable permissions.

If the value ofthisattribute is1 or 2, and a mount point doesnot exist,theagent createsa mountpoint with systemdefaultpermissionswhenthe resource is brought online. If the permissions of the mount pointis less than 555, a warning message is logged.

If the value of this attribute is 2, and the mount point does not exist,theagent createsa mountpoint with systemdefaultpermissionswhentheresource is brought online. If thepermissions for themountpointare less than 555, a warningmessage is logged. Inaddition,VCSdeletesthe mount point and any recursively created directories when theresource is brought offline. The mount point gets deleted only if it isempty, which is also true for recursive mount points.


Default: 0

CreateMntPt

This attribute specifies the group ownership of the mounted filesystem. The agent verifies the group ownership of the mounted filesystem every monitor cycle if the value of the AccessPermissionChkattribute is not 0.


Example: "grp1"

MntPtGroup

This attribute specifies theuser ownershipof themounted file system.The agent verifies the user ownership of the mounted file systemevery monitor cycleif thevalueof theAccessPermissionChk attributeis not 0.


Example: "usr1"

MntPtOwner

This attribute specifies the permissions of the mounted file systemin absolute format of a four-digit octal. The agent verifies the modeof the mounted file system every monitor cycle if the value of theAccessPermissionChk attribute is not 0.


Example: "0755"

MntPtPermission




static keylist RegList = { VxFSMountLock }static int IMF{} = { Mode = 3, MonitorFreq = 1,RegisterRetryLimit = 3 }

static str IMFRegList[] = { MountPoint, BlockDevice, FSType }static boolean AEPTimeout = 1static keylist SupportedActions = { "mountpoint.vfd","mounted.vfd", "vxfslic.vfd" , "chgmntlock", "mountentry.vfd" }static str ArgList[] = { MountPoint, BlockDevice, FSType,MountOpt, FsckOpt, SnapUmount, CkptUmount, OptCheck,CreateMntPt, MntPtPermission, MntPtOwner, MntPtGroup,AccessPermissionChk, RecursiveMnt, VxFSMountLock }str MountPointstr BlockDevicestr FSType

str MountOptstr FsckOptboolean SnapUmount = 0boolean CkptUmount = 1int OptCheck = 0int CreateMntPt = 0int ReuseMntPt = 0str MntPtPermissionstr MntPtOwnerstr MntPtGroup

int AccessPermissionChk = 0boolean RecursiveMnt = 0boolean VxFSMountLock = 0

)

Notes for Mount agentThe Mount agent has the following notes:

■ Support for spaces in directory names

■ Support for multiple bindfs


■ VxFS file system lock

■ IMF usage notes

■ Enabling Level two monitoring for the Mount agent




Support for spaces in directory namesTheMount agent supportsdirectorynameswith spaces. Thespacecanbe leading,trailing, or in the middle of the name. If the directory name has a trailing space,provide an extra "/"at theendof thecorresponding attribute ofa Mount resource.The attributes that currently support spaces in directory names are MountPointand BlockDevice. Note that the agent does not support spaces created using theTAB key.

Support for multiple bindfsThe Mount agent supports file system of the type bind. It also allows multiplemounts of type bind from the same block device. However, the mount points aredifferent. So, with the FSType attribute set to bind, the mount resource does not

reportUNKNOWN state when thesameblock device is mountedon anothermountpoint.

Sample configuration

Mount bindmount (MountPoint = "/bind"BlockDevice = "/test/bind"FSType = bindMountOpt = rw)

Mount bindmount1 (MountPoint = "/bind1"BlockDevice = "/test/bind"FSType = bindMountOpt = rw)

Mount mount (

MountPoint = "/test"BlockDevice = "/dev/vx/dsk/testdg/testvol"FSType = vxfsMountOpt = rwFsckOpt = "-y")

bindmount requires mount


74



bindmount1 requires mount

High availability fire drillThehigh availability firedrill detects discrepancies between theVCSconfigurationand the underlying infrastructure on a node; discrepancies that might prevent aservice group from going online on a specific node.

For Mount resources, the high availability drill performs the following, it:

■ Checks if the specified mount point directory exists

■ Checks if the mount point directory is already used

■ Checks for valid Veritas (VxFS) file system licenses

■ Checks if the mount point exists in the /etc/fstab file

For more information about using the high availability fire drill, see the VeritasCluster Server Administrator ’ s Guide .

VxFS file system lockIf the mount option in the mount table output has the option mntlock=" key ", thenit is locked with the key " key ". To verify if mount locking is in use and has thevalue of " key ", run the mount command and review its output.

# mount

If the VxFS file system has mntlock=" key " in itsmount options, then unmountingthe file system fails.

You can unlock the file system with the fsadm command and then unmount it. Tounlock a locked mount, run the following command where " key " is the lockidentifier and mount_point_name is the file system mount point.

# /opt/VRTS/bin/fsadm -o mntunlock=" key " mount_point_name

To unmount a file system mounted with locking, run the vxumount command withthe option mntunlock=" key ", for example:

# /opt/VRTS/bin/vxumount -o mntunlock=" key " mount_point_name




■ # haconf -dump -makero

For more details about the LevelTwoMonitorFreq attribute, refer to the VeritasCluster Server Agent Developer ’ s Guide .

Sample configurations for Mount agent

Basic SCSI configuration for Mount agentConfiguration for Linux follows:

Mount mnt_r1_1 (MountPoint = "/testdir/LVM_R1_1"BlockDevice = "/dev/mapper/emc_vg-emc_r1"

FSType = ext3MountOpt = "rw"FsckOpt = "-y")

VxFS mount lock example for Mount agent

Mount test_mnt (MountPoint = "/home/export"BlockDevice = "/dev/vx/dsk/nfsdg/vol0"FSType = vxfs

MountOpt = rwFsckOpt = "-n"VxFSMountLock = 1)

NFS mount example for Mount agent

Mount mnt1 (MountPoint = "/mnt/vcstest"BlockDevice = "sys1:/home/export"

FSType = nfs)

Debug log levels for Mount agentThe Mount agent uses the following debug log levels:

DBG_1, DBG_2, DBG_3, DBG_4, DBG_5





78



High availability fire drill for IP agentThehigh availability firedrill detects discrepancies between theVCSconfigurationand the underlying infrastructure on a node. These discrepancies might preventa service group from going online on a specific node. For IP resources, the highavailability fire drill:

■ Checks for the existence of a route to the IP from the specified NIC

■ Checks for the existence of the interface configured in the IP resource


Dependencies for IP agentIP resources depend on NIC resources.

Figure 3-1 Sample service group that includes an IP resource

Application

IP Mount

NIC VolumeSet

DiskGroup

Agent functions for IP agent

Configuresthe IP address to the NIC. Checks ifanother system isusingthe configured IP address and issues a warning. For IPv4 addresses,it uses the ifconfig command to set the IPv4 address on an uniquealias of the interface if the Options attribute is configured and none

of the IPOptions and IPRouteOptions attributes are configured, elseit makes use of the ip command. For IPv6 addresses, the ip commandis used.

Online

■ route.vfdChecksfor theexistence of a route to theIP from thespecified NIC.

■ device.vfdChecks for the existence of the interface configured in the Deviceattribute.

Action

Network agentsIP agent



Brings down the IP address that is specified in the Address attribute.Offline

Monitors the interface to test if the IP address that is associated withthe interface is alive.

Monitor

Brings down the IP address that is specified in the Address attribute.Clean

State definitions for IP agentThe state definitions for this agent follow:

Indicates that thedevice is up andthe specifiedIP address is assignedto the device.

ONLINE

Indicates that the device is down or the specified IP address is notassigned to the device.

OFFLINE

Indicates that the agent could notdetermine thestate of theresourceor that the resource attributes are invalid.

UNKNOWN

Indicates that the IP address could not be brought online, usuallybecause the NIC configured in the IP resource is faulted or the IPaddress was removed out of VCS control.

FAULTED

Attributes for IP agentFor Linux:


82



Table 3-1 Required attributes

DescriptionRequired attribute

A virtual IP address that is associated with the interface, andwhich is different from the base IP address.


Examples:

■ IPv4:"192.203.47.61"

■ IPv6:"2001::10"

Note: For configuring the IP address of a different networkthan the network of the base IP address, you need to configurethe following values in /etc/sysctl.conf.

File: /etc/sysctl.conf

# avoid deleting secondary IPs on deleting theprimary IP

net.ipv4.conf.default.promote_secondaries = 1

net.ipv4.conf.all.promote_secondaries = 1

After changing these values, load the current values of the /etc/ sysctl.conf file using the following command: # sysctl

-p /etc/sysctl.confThis is to make sure that if the primary IP address isunplumbed, thesecondary IPaddress fora network ispromotedto the primary address.

Address

The name of the NIC device that is associated with the IPaddress. Requires the device name without an alias.


Example: eth0

In above example, eth0 is specified to assign the IP address to

the next available alias of eth0.

Use the ip addr command to display a list of NICs that areup and the IP addresses assigned to each NIC.

Device

See Table 3-2 on page 84.One of the followingattribute:

■ PrefixLen■ NetMask






The subnetmask that is associated with theIP address. For theIPv4 protocol, specify thevalue of NetMask attribute indecimal(base 10).

Configure this attribute if the IP address is an IPv4 address.


Example: "255.255.255.0"

NetMask


DescriptionOptional

attribute

This is the prefix for the IPv6 address represented as the CIDR value.

Type-dimension: integer-scalar

Range: 0 - 128

Default: 1000

Note: Note that the default value is intentionally invalid for thisattribute. You must set the value of this attribute to a range from 0 to128 to activate this attribute.

Example: 64

PrefixLen

Options for the ifconfig command.

The ifconfig command generally resembles:

ifconfig dev inet ipv4addr netmask netmask Optionsup

Youmust configure either this Optionsattribute or boththe IPOptionsand IPRouteOptions attributes.

Do not configure any options in this attribute that you can specify inother attributes. For example, for the netmask use the NetMask

attribute.

For complete list of ifconfig options refer to ifconfig manpage .


Example: "broadcast 172.20.9.255"

Options


84





Specifies the extra options that are passed to the ip addr addcommand.

The agent uses this attribute in tandem with the IPRouteOptionsattribute.

The ip addr add command generally resembles:

"ip -4 addr add ipv4addr/prefixlen IPOptions labellabel dev device "

Note: If you configure this attribute, the agent ignores the Optionsattribute and uses the ip command instead of ifconfig .


Examples:

■ "broadcast 172.20.9.255"■ "scope link"

IPOptions

Specifies the extra options that are passed to the ip route addcommand.

The agent uses this attribute in tandem with the IPOptions attribute.

Theip route add

command resembles:

"ip route add IPRouteOptions dev device "

Note: If you configure this attribute, the agent ignores the Optionsattribute and uses the ip command instead of ifconfig.


Examples:

■ "default via 172.20.9.1"■ "scope link"

IPRouteOptions

Resource type definition for IP agentThe resource definition for this agent on Linux follows:

type IP (static keylist RegList = { NetMask }static keylist SupportedActions = { "device.vfd", "route.vfd" }static str ArgList[] = { Device, Address, NetMask, PrefixLen,




Options, IPOptions, IPRouteOptions }str Devicestr Address

str NetMaskint PrefixLen = 1000str Optionsstr IPOptionsstr IPRouteOptions

)

Sample configurations for IP agentThe sample configurations for this agent follow:

IPv4 ConfigurationConfiguration 1 for Linux follows:

IP IP_192_203_47_61 (Device = eth0Address = "192.203.47.61"NetMask = "255.255.248.0")

IPv6 ConfigurationConfiguration using a specified NetMask for Linux follows:

IP IP_2001_10 (Device = eth0Address = "2001::10"PrefixLen = 64)

Debug log levels for IP agentThe IP agent uses the following debug log levels:DBG_1, DBG_2, DBG_4, DBG_5

NIC agentThe NIC agent monitors the configured NIC. If a network link fails, or if a problemarises with the NIC, the resource is marked FAULTED. You can use the agent to

Network agentsNIC agent

86



Agent functions for NIC agent

■ If the NICmaintains its connection status, theagent uses

MII to determine the status of the resource.If the NIC does not maintain its connection status, theagent verifies that theNICis configured. The agent thensends a ping to all the hosts that are listed in theNetworkHosts attribute. If the ping test is successful, itmarks the NIC resource ONLINE.If the NetworkHosts attribute list is empty, or the pingtest fails, the agent counts the number of packets thatthe NIC received. The agent compares the count with apreviously stored value. If the packet count increases,the resource is marked ONLINE. If the count remains

unchanged, the agent sends a ping to the broadcastaddress of the device to generate traffic on thenetwork.The agent counts the number of packets that the NICreceives before and after the broadcast. If the countincreases, the resource is marked ONLINE. If the countremains the same or decreases over a period of fivebroadcast cycles, the resource faults.

Monitor

■ device.vfdChecks for the existence of the interface configured inthe Device attribute.

Action

Note: The NIC agent supports ethtool and MII-based device status monitoring. If the Mii attribute for a NIC resource is set to 1, the agent first checks for the devicestatus using ethtool and if it fails, it checks for the MII status for the device. Theagent does notdo anyping andpacketcount-based monitoring if it cansuccessfullydecide the status of the NIC device based on ethtool and MII tests.

State definitions for NIC agentThe state definitions for this agent follow:

Indicates that the NIC resource is working.ONLINE

Indicates that the NIC has failed.FAULTED

Indicates the agent cannot determine the interface state. It may bedue to an incorrect configuration.

UNKNOWN


88



Attributes for NIC agentTable 3-3 Required attributes


Specifies the name of the NIC that you want to monitor.

Use the ip addr command to list all network adapters and the IPaddresses assigned to each NIC.


Example: "eth0" or "eth1"

Device



Flag that defines whether the NIC maintains its connection status.

If this flag is set to 1, the agent uses ethtool and MII hardwareregisters, instead of the ping and packet count method. The agentuses this method to determine the health of the network card.

If the flag is set to 0, the agent doesnot use ethtool and Mii to monitorthe status of the NIC.

Type and dimension: integer-scalarDefault: 1

Mii






List of hosts on the network that are pinged to determine the state of the NIC. Specify the IP address of the host — not the host name.

The specified hosts must be pingable:

■ from all the cluster nodes that are specified in the SystemListattribute for the service group to which the resource belongs

■ through the device that is specified in the Device attribute

The command to ping the host (hostip) via a NIC device (nicdev) is:

For IPv4:

# ping -I device hostipFor IPv6:

# ping6 -I device hostip

If more than one network host is listed, the monitor returns ONLINEif the ping test is successful with at least one of the hosts.

You can use both IPv4 and IPv6 NetworkHost addresses, and you canconfigure both types of addresses in the same resource.


Example:

IPv4:

{ "166.93.2.1", "166.99.1.2" }

IPv6:

{ "2001::1" , "166.93.2.1" }

NetworkHosts

Attribute that defines whether theagent sendsa broadcast ping beforeit retrieves the received packet statistics. This attribute is used whenMii is not set and no network hosts are specified.

If the value of this attribute is 1, the agent retrieves received packet

statistics fromthenetstatcommandandcompare themwith previouslystored values. The agent sends a broadcast ping to the network onlyif the packet count remains unchanged.

If the value of this attribute is 0, the agent sends a broadcast pingbefore it checks the network statistics.


Default: 1

PingOptimize


90



Resource type definition for NIC agentThe resource definition for this agent on Linux follows:

type NIC (static keylist SupportedActions = { "device.vfd" }static int OfflineMonitorInterval = 60static str ArgList[] = { Device, PingOptimize, Mii, NetworkHosts}static str Operations = Nonestr Deviceint PingOptimize = 1int Mii = 1str NetworkHosts[])

Monitoring bonded NICs for NIC agentThe NIC agent can monitor the network interfaces (bond0, bond1, etc.) that thebonding driver exports. Refer to operating system vendor documentation to setup the bonds and to configure your system to load the bonding driver correctly.

For monitoring a bond interface, the two important settings are:

■ The value of the miimon parameter, which you set while loading the bondingdriver. miimon is a parameter to the bonding module and has a default setting

of 0.■ The value of the Mii attribute (Mii) of the NIC resource, which you set at

runtime. Mii is an attribute of the NIC resource and has a default setting of 1.

Setting Mii and miimon for NIC agentFor the following cases, the name of the monitored bond interface is B. If you donot use one of the following cases to set up bonding, the bonding driver canpotentially provide incorrect health status. This incorrect health status can resultin VCS failing to fault the resource appropriately.

Case 1Accept defaults — miimon is 0 and Mii is 1. Each of B ’ s slaves must support thenetif_carrier_ok in-kernel call.




Figure 3-3 Sample service group that includes an IPMultiNIC resource

Application

IPMultiNIC Mount

MultiNICA Volume

DiskGroup

Agent functions for IPMultiNIC agentConfigures a virtualIP address on theactiveinterface of theMultiNICAresource. Also sends out a gratuitous ARP.

Online

Removes the virtual IP address from the active interface of theMultiNICA resource.

Offline

Checks if the virtual IP address is configured on one interface of theMultiNICA resource.

Monitor

Removes the virtual IP address from active interface of the MultiNICAresource.

Clean

State definitions for IPMultiNIC agentThe state definitions for this agent follow:

Indicates that the specified IP address is assigned to one of theinterfaces specified in the corresponding MultiNICA resource.

ONLINE

Indicatesthat thespecified IP address is notassigned to anyinterfaceof the MultiNICA resource.

OFFLINE

Indicates that the agent can not determine the state of the resource.This state may be due to an incorrect configuration.

UNKNOWN

Indicates that the IP address could not be brought online, usuallybecause all the NICs in the MultiNICA resource are faulted or the IPaddress was removed out of VCS control.

FAULTED

Network agentsIPMultiNIC agent

94



Attributes for IPMultiNIC agentTable 3-5 Required attributes


The virtual IP address that is assigned to the active NIC.


Examples:

■ IPv4: "10.128.10.14"■ IPv6: "2001:DB8::"

Address

Nameof theassociatedMultiNICA resourcethat determinestheactive

NIC.Type and dimension: string-scalar

Example: "mnic"

MultiNICAResName

See Table 3-6 on page 95.One of the twoattributes:

■ NetMask■ PrefixLen



The ifconfig command options for the virtual IP address. Do notconfigure any options in this attribute that you can specify in otherattributes. For example, for the netmask use the NetMask attribute.


Example: "mtu 2000"

Options






Specifies the extra options that are passed to the ip addr add

command. The ip addr add command resembles the following:

■ IPv4"ip addr add ipv4addr/prefixlen IPOptions labellabel dev device "

■ IPv6"ip addr add ipv6addr/prefixlen IPOptions labellabel dev device "


Note: If you configure this attribute, the agent ignores the Optionsattribute and uses the ip command instead of ifconfig.

Example:


IPOptions

FortheIPv4 protocol, specify thevalueofNetMask attributeindecimal(base 10).


Type and dimension: string-scalarExample: "255.255.255.0"

NetMask

Specifies theprefixfortheIPv6address represented as theCIDR value.When you use the IPv6 protocol, you must configure a value for thisattribute.


Range: 0 - 128

Default: 1000

Note: The default value is intentionally invalid for this attribute. Youmust set the value of this attribute to a range from0 to 128 toactivatethis attribute.

Example: 64

PrefixLen

Resource type definition for IPMultiNIC agentThe resource definition for this agent on Linux follows:


96



type IPMultiNIC (static int MonitorTimeout = 200static int OfflineMonitorInterval = 120

static int ToleranceLimit = 2static str ArgList[] = { Address, NetMask, PrefixLen,MultiNICAResName, Options, IPOptions, "MultiNICAResName:Probed"}str Addressstr MultiNICAResNamestr NetMaskint PrefixLen = 1000str Optionsstr IPOptions)

Sample configuration: IPMultiNIC and MultiNICARefer to the MultiNICA agent for more information.

IPv4 configuration for IPMultiNIC agent

cluster foo (UserNames = { admin = "cDRpdxPmHpzS." }

CounterInterval = 5)system sysA ()system sysB ()group grp1 (SystemList = { sysA = 1, sysB = 2 })IPMultiNIC ip1 (Address = "192.123.10.177"MultiNICAResName = mnicNetMask = "255.255.248.0")MultiNICA mnic (Device @sysA = { eth0 = "192.123.10.127", eth1 ="192.123.11.127" }Device @sysB = { eth0 = "192.123.10.128", eth2 ="192.123.11.128" }




NetMask = "255.255.248.0"NetworkHosts = { "192.123.10.129", "192.123.10.130" })

ip1 requires mnic// resource dependency tree//// group grp1// {// IPMultiNIC ip1// {// MultiNICA mnic// }// }

IPv6 configuration for IPMultiNIC agent

cluster foo (UserNames = { admin = "cDRpdxPmHpzS." }CounterInterval = 5)system sysA ()system sysB ()

group grp1 (SystemList = { sysA = 1, sysB = 2 })IPMultiNIC ip1 (Address = "2001::110"MultiNICAResName = mnicPrefixLen=96)MultiNICA mnic (Device @sysA = { eth0 = "2001::10", eth1 = "2001::11" }Device @sysB = { eth0 = "2001::12", eth2 = "2001::13" }PrefixLen=96NetworkHosts = { "2001::1", "2001::500" })ip1 requires mnic// resource dependency tree//// group grp1// {


98



// IPMultiNIC ip1// {// MultiNICA mnic

// }// }

Mixed mode configuration —IPv4 and IPv6 for IPMultiNIC agentMixed mode configuration for IPv4 and IPv6 follows:

cluster foo (UserNames = { admin = "cDRpdxPmHpzS." }

CounterInterval = 5)system sysA ()system sysB ()group grp1 (SystemList = { sysA = 1, sysB = 2 })IPMultiNIC ip1 (Address = "2001::110"

MultiNICAResName = mnicPrefixLen=96)IPMultiNIC ip2 (Address = "192.123.10.177"MultiNICAResName = mnicNetMask="255.255.248.0")MultiNICA mnic (Device @sysA = { eth0 = "192.123.10.127", eth1 =

"192.123.11.128" }Device @sysB = { eth0 = "192.123.10.129", eth2 ="192.123.11.130" }NetMask = "255.255.248.0"DualDevice @sysA = { eth0 = "2001::10", eth1 ="2001::11" }DualDevice @sysB = { eth0 = "2001::12", eth2 ="2001::13" }PrefixLen=96




NetworkHosts = { "2001::1", "192.123.10.129" })

ip1 requires mnicip2 requires mnic// resource dependency tree//// group grp1// {// IPMultiNIC ip1// {// MultiNICA mnic// }// IPMultiNIC ip2

// {// MultiNICA mnic

// }// }

Debug log levelsThe IPMultiNIC agent uses the following debug log levels:

DBG_1, DBG_2, DBG_4, DBG_5

MultiNICA agentThe MultiNICA represents a set of network interfaces, and provides failovercapabilities between them. You can use the agent to make IP addresses onmultiple-adapter systems highly available.

The IPMultiNIC agent depends upon the MultiNICA agent to select the mostpreferred NIC on the system. IPMultiNIC brings the virtual IP online or offline.

However, if theMultiNICA resource changes its activedevice, theMultiNICA agenthandles the shifting of IP addresses.

If a NIC on a system fails, the MultiNICA agent selects another active NIC. Theagent then shifts the virtual IP address to the newly selected active NIC. Only ina case where all the NICs that form a MultiNICA agent fail, does the virtual IPaddress shift to another system.

If you associate an interface with a MultiNICA resource, do not associate it withany other MultiNICA or NIC resource.

Network agentsMultiNICA agent

100



If the same set of interfaces must be a part of multiple service groups, configure:

■ A MultiNICA resource in one of the service groups, and

■

The Proxy resources that point to the MultiNICA resource in the other servicegroups.

The MultiNICA agent can operate in two modes:

■ IP Conservation Mode (ICM) for MultiNICA agent

■ Performance Mode (PM) for MultiNICA agent

With sufficient IP addresses, use PM.

The IPMultiNIC and MultiNICA agents support Linux bonds.

The IPMultiNIC and MultiNICA agents support IPv4 and IPv6.

Dependencies for MultiNICA agentThe MultiNICA resource does not depend on any other resources.

Figure 3-4 Sample service group that includes a MultiNICA resource

Application

IPMultiNIC Mount

MultiNICA Volume

DiskGroup

IP Conservation Mode (ICM) for MultiNICA agentRequires fewer IP addresses thanPerformanceMode, but provides slower failover.

Configuration for MultiNICA agentWhen a MultiNICA resource is to be configured in ICM, the configured NICs musthave the samebaseIPaddress.This IPaddress must beunique, and cannot appearon any other NIC on any other node. You do not need to enable the base IPaddresses beforehand. This mode does not support failing back the NIC, see theoptional Failback attribute.




Operation for MultiNICA agentWhen you specify all the NICs with the same base IP address, the agent runs inICM. It enables the base IP address on the active NIC.

In case of a failover, it moves the base IP address to the new active NIC. It alsomoves all the virtual IP addresses that are configured on that NIC. It tries to findthe next working NIC in the order of priority.

Performance Mode (PM) for MultiNICA agentRequires more IP addresses than ICM, but provides faster failover. You do nothave to spend time enabling and disabling base IP addresses and reinstating lostroutes, thus no resultant service disruption occurs.

Configuration for MultiNICA agentWhen the MultiNICA resource is to be configured in PM, each NIC must have aunique base IP address. The base IP address cannot appear on any other NIC onthe same node or any other node. The base IP address of all the devices in a singleMultiNICA resource must belong to the same subnet in the configuration.

When you configure a single NIC under a MultiNICA resource, the MultiNICAagent uses PM. The base IP addresses have to be enabled on each NIC underMultiNICA control before starting VCS and handing over the management of theNICs to the agent.

Operation for MultiNICA agentThe agent uses this modewhen all NICs under the MultiNICA agent haveseparatebase IP addresses specified.

The mode requires that you enable the base IP addresses before starting VCS.When a NIC goes down, the agent migrates only virtual IP addresses.

In this mode, you can set the Failback attribute to 1 or 0:

■ If you set the Failback attribute to 1, in each monitor cycle the agent checksto see if a preferred NIC is up. If the NIC is up, it selects that NIC as the activeNIC and moves the virtual IP addresses to the preferred NIC.

■ If you set the Failback attribute to 0, the agent selects a new active NIC onlyif the current active NIC fails. It selects the new active NIC in the order of priority.

Agent function for MultiNICA agentThe agent functions for this agent follow:


102



Uses ethtool and Media Independent Interface (MII) to request thedevice status. If the hardware does not respond, the agent sends aping to the hosts that are listed in the NetworkHosts attribute. If the

ping test fails, the agent checks for activity on a configured interfaceby sampling the input packets that are received on that interface. If the agent does not detect activity, it forces activity by sending out abroadcast ping. If the agent does not receive a network reply, itmigrates to the most suitable next interface.

Monitor

Attributes for MultiNICA agentWhile configuring the MultiNICA resource, in addition to the required attributes,you must also configure at least one set of attributes from the following:

■

For IPv4 configurations:■ The Options attribute and the RouteOptions attribute, or

■ The IPv4RouteOptions attribute

■ For IPv6 configurations:

■ The IPv6RouteOptions attribute




Table 3-7 Required attributes


List of devices and associated base IP addresses. This attribute mustbe specified separately for each system in the SystemList. You mustspecify the devices in the list in the order of priority. The first devicethat theagent determines as "up" becomes theactive device, to whichthe agent assigns a corresponding IP address.

ForIP ConservationMode(ICM): ifall theNICsconfigured in theDeviceattribute are down, the MultiNICA agent faults the resource after a2-3 minute interval. This delay occurs because the MultiNICA agenttests the failedNICseveral times before it marks theresource offline.The engine log records messages that provide a detailed description

of the failover events. Find the engine log in /var/VRTSvcs/log/engine_A.log.

For each system you must localize the attribute with a separate baseIP address.

Type and dimension: string-association

Examples:

■ IPv4Device@vcsonelnx1={ eth1 = "10.212.100.178", eth2 ="10.212.100.179" }Device@vcsonelnx2 ={ eth2 = "10.212.100.180", eth3 ="10.212.100.181" }

■ IPv6Device@vcsonelnx1={ eth1 = "1234::5678", eth2 = "1234::5679" }Device@vcsonelnx2={ eth3 = "1234::5680", eth4 = "1234::5681"}

You can use IPv4 or IPv6 base addresses in the Device attribute, butmake sure all of the base addresses use a common IP version.

Device

One of the twoattributes:

■ NetMask■ PrefixLen


104





The DualDevice attribute specifies the list of devices and associated IPv6 base addresses.Specify:

■ this attribute separately for each system in the SystemList.■ the devices in the list in the order of priority.

The first device that the agent determines as "up" becomes the active device, to whichthe agent assigns a corresponding IP address.

NICs in Device and DualDevice attributes should be identical and in the same order.

Usethe DualDevice attributeonlywhen configuring mixed IPv4/IPv6stacks. In that case,use the Device attribute to configure the IPv4 stack and the DualDevice attribute to

configure the IPv6 stack.Example:

DualDevice@vcslnx1={ eth1 = 2001::DB8, eth2 = 2001::DB9}

DualDevice@vcslnx2={ eth3 = 2001::DB10, eth4 = 2001::DB11}

DualDevice

This attribute determines if the active NIC should be changed to a preferred NIC, eventhough the current NIC is healthy. If operating in the ICM mode, change the value to 0.


Default: 1

Failback

The extra options that are passed to the ip addr add command for IPv4 addresses..

The agent uses this attribute in tandem with the IPv4RouteOptions attribute.

The ip addr add command generally resembles:

"ip addr add ipv4addr/prefixlenIPv4AddrOptions dev device "

Note: If you configure this attribute, the agent ignores the Options attribute and usesthe ip command instead of ifconfig.


Example:


IPv4AddrOptions






Specifies options for the ip link command, which can bring an interface up or down.The ip link command generally resembles:

"ip link dev up LinkOptions "

Note: If you configure this attribute, the agent ignores the Options attribute and usesthe ip command instead of ifconfig.


Example:

■ "mtu 1024"■ "broadcast 172.20.9.255"

LinkOptions

List of hosts on the network that receive pings to determine the state of the NICs. Specifythe IP address of the host, not the host name. Include the hosts that all the NICs in theDevice list can reach. If more than one network host is listed, monitor returns ONLINEif the ping test is successful with at least one of the hosts.

You canuse both IPv4 and IPv6 NetworkHost addresses, andyou canconfigure both typesof addresses in the same resource.


Example: NetworkHosts = { "2001::1", "192.123.10.129" }

NetworkHosts

The ifconfig options that you want to use when you assign the base IP address to theactive device.

You must configure either this Options attribute or both the IPv4AddrOptions andIPv4RouteOptions attributes.

This attribute does not support IPv6.

Note: If you configure any of the IPv4AddrOptions or LinkOptions attribute, the agentignores Options attribute and uses ip command instead of ifconfig.


Example: "broadcast 10.212.100.255"

Options

Determines whether or not a broadcast ping is sent before checking network statistics,which are used to determine the state of the NIC (if MII is not supported and the ping toNetworkHosts does not confirm the NIC is up.) A value of 1 indicates a broadcast pingdoes not occur, a value of 0 indicates a broadcast ping occurs.


Default: 1

PingOptimize






Assignment of a base IP address to a device followed by a route add command. Thecommand takes the options specified by this attribute. RouteOptionsare applicable onlywhen configuring the local host as the default gateway. No routes are added if this stringis set to NULL.

The RouteOptions attribute works in tandem with the Options attribute. You mustconfigure the Options attribute when you use this attribute or the agent ignores thisattribute.


Example: "default gw 166.98.16.103""

RouteOptions

Specifies the prefix for the IPv6 address represented as the CIDR value.When you use the IPv6 protocol, you must configure a value for this attribute.


Range: 0 - 128 Default: 1000

Note: The default value is intentionally invalid for this attribute. Youmust set the valueof this attribute to a range from 0 to 128 to activate this attribute.

Example: 64

PrefixLen

Specifies the netmask that is associated with the base IP address. The value must bespecified in decimal (base 10).



Example: "255.255.252.0"

NetMask

Flag that defines whether the NIC maintains its connection status.

If this flag is set to 1, the agent uses ethtool and MII hardware registers, instead of theping and packet count method. The agent uses this method to determine the health of the network card.

If the flag is set to 0, the agent does not use ethtool and Mii to monitor the status of the

NIC.Type and dimension: boolean-scalar

Default: 1

Mii


108



Resource type definition for MultiNICA agenttype MultiNICA (

static int MonitorTimeout = 240static str ArgList[] = { Device, DualDevice, NetMask,PrefixLen, Options, RouteOptions, PingOptimize,MonitorOnly, NetworkHosts, Failback, LinkOptions,IPv4AddrOptions, IPv6AddrOptions, IPv4RouteOptions,IPv6RouteOptions, Mii }static str Operations = Nonestr Device{}str DualDevice{}str NetMaskint PrefixLen = 1000

str Optionsstr RouteOptionsstr LinkOptionsstr IPv4AddrOptionsstr IPv6AddrOptionsstr IPv4RouteOptionsstr IPv6RouteOptionsint PingOptimize = 1str NetworkHosts[]boolean Failback = 1

boolean Mii = 1)

Sample configurations for MultiNICA agentThe sample configurations for the following agent are:

MultiNICA and IPMultiNIC Performance Mode configurationIn this example, two systems (sysA and sysB) eachhavea pairofnetwork interfaces(eth0 and eth1, eth0 and eth2). These interfaces have different physical IP

addresses and the agent behaves in Performance Mode (PM).The MultiNICA resource fails over only the logical IP address to the backup NICin the event ofa failure. The resource ip1 has theAddress attribute, which containsthe logical IP address. In the event of a NIC failure on sysA, the logical IP addressfails over from eth0 to eth1. In the event that eth1 fails — the address fails back toeth0 — as long as eth0 is reconnected.




However, if both theNICs on sysA aredisconnected, theMultiNICA andIPMultiNICresources work in tandem to fault the group on sysA. The entire group fails overto sysB.

Ifyouhave more than oneservice groupusing theMultiNICA resource, thesecondservice groupcanusea Proxy resource.TheProxy resourcepoints to theMultiNICAresource of the first service group. This resource prevents redundant monitoringof the NICs on the same system. The IPMultiNIC resource is always madedependent on the MultiNICA resource.

cluster foo (UserNames = { admin = "cDRpdxPmHpzS." }CounterInterval = 5)

system sysA ()system sysB ()

group grp1 (SystemList = { sysA = 1, sysB = 2 })

IPMultiNIC ip1 (

Address = "192.123.10.177"MultiNICAResName = mnicNetMask = "255.255.248.0")

MultiNICA mnic (Device @sysA = { eth0 = "192.123.10.127", eth1 ="192.123.11.128" }Device @sysB = { eth0 = "192.123.10.129", eth2 ="192.123.11.130" }NetMask = "255.255.248.0"NetworkHosts = { "192.123.10.1", "192.123.10.2" })

ip1 requires mnic// resource dependency tree//// group grp1// {


110



MultiNICAResName = mnicNetMask = "255.255.248.0")

MultiNICA mnic (Device @sysA = { eth0 = "192.123.10.127", eth1 ="192.123.10.127" }Device @sysB = { eth0 = "192.123.10.128", eth2 ="192.123.10.128" }NetMask = "255.255.248.0"NetworkHosts = { "192.123.10.1", "192.123.10.2" }Failback = 0

)

ip1 requires mnic

// resource dependency tree//// group grp1// {// IPMultiNIC ip1// {// MultiNICA mnic

// }// }

IPv6 configuration for MultiNICA agentThe following sample is for IPv6 use.

group mnica_group (SystemList = { sysA = 0, sysB = 1 })

IPMultiNIC ipmnic_res (Address = "2007:192::1627:161"MultiNICAResName = mnica_resPrefixLen = 64)

MultiNICA mnica_res (Device @sysA = { eth0 = "fe80::214:4fff:fe96:ae0a",


112



eth1 = "fe80::214:4fff:fe96:ae0a" }Device @sysB = { eth0 = "fe80::214:4fff:fe96:ae0b",eth2 = "fe80::214:4fff:fe96:ae0b" }

Failback = 0PrefixLen = 64)

ipmnic_res requires mnica_res

Mixed mode configuration —IPv4 and IPv6 for MultiNICA agentMixed mode configuration for IPv4 and IPv6 follows:

cluster foo (UserNames = { admin = "cDRpdxPmHpzS." }CounterInterval = 5)system sysA ()system sysB ()group grp1 (SystemList = { sysA = 1, sysB = 2 })IPMultiNIC ip1 (Address = "2001::110"MultiNICAResName = mnicFailback = 0PrefixLen=96)IPMultiNIC ip2 (Address = "192.123.10.177"MultiNICAResName = mnicNetMask="255.255.248.0"

)MultiNICA mnic (Device @sysA = { eth0 = "192.123.10.127", eth1 ="192.123.11.127" }Device @sysB = { eth0 = "192.123.10.128", eth1 ="192.123.11.128" }NetMask = "255.255.248.0"DualDevice @sysA = { eth0 = "2001::10", eth1 = "2001::10" }




DualDevice @sysB = { eth0 = "2001::11", eth1 = "2001::11" }Failback=0PrefixLen=96

NetworkHosts = { "2001::1", "192.123.10.1" })ip1 requires mnicip2 requires mnic// resource dependency tree//// group grp1// {// IPMultiNIC ip1// {// MultiNICA mnic

// }// IPMultiNIC ip2// {// MultiNICA mnic// }// }

Debug log levels for MultiNICA agentThe MultiNICA agent uses the following debug log levels: DBG_1, DBG_2, DBG_3,

DBG_4, DBG_5

DNS agentThe DNS agent updates and monitors the mapping for the following:

■ The host name to IP address (A, AAAA, or PTR record)

■ Alias to hostname or canonical name (CNAME)

The agent performs these tasks for a DNS zone when failing over nodes acrosssubnets (a wide-area failover). Resource records (RR) can include different types:A, AAAA, CNAME, and PTR records.

Use the DNS agent if the Resource Records need to be dynamically added anddeleted from the DNS servers during failover. The agent updates the name serverwith the new resource record mappings while failing over and allows the clientsto connect to the failed over instance of the application.

For important information about this agent, refer to Agent notes for DNS agent

Network agentsDNS agent

114



Dependencies for DNS agentNo dependencies exist for the DNS resource.

Figure 3-5 Sample service group that includes a DNS resource

DB

IP Mount

NIC Volume

DiskGroup

DNS

Agent functions for DNS agent

Updates one or more name servers with the resource records.

The agent updates the name servers defined in the StealthMasters

attribute. If you have not configured this attribute then the agentobtains the name of the master server by sending anStart of Authority(SOA) query. This query retrieves the SOA record of the zone definedin the agent's Domain attribute. This SOA record contains the nameof the master server.

The agent creates PTR records for each RR of type A or AAAA if thevalue of the CreatePTRattribute is true.A prerequisite for this featureis that the same master or stealth server serves the forward (A orAAAA) and reverse zones.

Finally the agent generates an Online lock file to indicate that theresource is online on the current system.

Note: The DNS agent does not send any update for a resource recordif it is already present on the name server.

Online

Removes the Online lock file.

If attribute OffDelRR is true, offline removes all records that theResRecord keys define.

Offline




Returns the ONLINE state if at least one name server reports allmappings that ResRecord defines. The name servers are the masteror StealthMaster servers and all the servers for which an NS record

for the zone exists.The monitor entry point also sends periodic refresh requests to DNSserver if the RefreshInterval attribute is set.

Monitor

Removes the Online lock file, if it exists. If attribute OffDelRR is true,clean removes all records that the ResRecord keys define.

Clean

Removes the Online lock file if the resource is reported online onanother node inside the cluster to prevent concurrency violation. If the lock file exists, at least one name server has to report all therecords that the ResRecord attribute defines. If all the name serversfail to report all the records, the agent function removes the Onlinelock file.

Open


■ keyfile.vfdThis action entry point checks if the key file as specified in theTSIGKeyFile attribute exists either locally or on shared storage.

■ dig.vfdThis action entry point checks if dig and nsupdate binaries existand are executable.

■ master.vfdThis action entry point checks if stealth masters are able to replyto SOA query for the configured domain.

Action

State definitions for DNS agentThe state definitions for this agent follow:

Online lock file exists and at least one name server can return allconfigured resource records.

ONLINE

At least one of the following is true:

■ The online lock does not exist.■ None of the name servers can report all of the RRs ’ mappings.

OFFLINE

Indicates that the DNS resource is not configured correctly. Canindicate that the resource record list contains an invalid value as apart of the record key or a record value of the ResRecord attribute.

UNKNOWN


116



Attributes for DNS agentTable 3-9 Required attributes


A string representingthe DNSzone that theagentadministers.

The domain name can only contain alphanumeric symbolsand the dash.


Examples:

■ Forward mapping: "demo.example.com"■ IPv4 reverse mapping: "2.168.192.in-addr.arpa"

Domain






ResRecord


118





follows:■ PTR: if the Domain attribute ends with .arpa■ A: if the record data field is an IPv4 address (four sets of

numbers, where a periodseparates each set. The followingdetails the pattern it tries to match:[1-223 ].[0-255 ].[0-255 ].[0-255 ] Hexadecimal is notsupported.)

■ AAAA: if the record data fields are in multiple sets of hexadecimal format, then this recordis an IPv6 associatedtype AAAA record.

■ CNAME: for any other valid record data.

Note: If a name in the ResRecord attribute does not complywith RFC 1035, then the agent logs a warning message to theengine log file. This ResRecord association is not used. As anexception to this, the DNS agent allows underscore character("_") in hostnames. Make sure that the DNS server supportsthe underscore character before you configure any DNSresource records to have the underscore character in theirhostnames.



This attribute (a non-zero integer) represents the Time To Live (TTL)value, in seconds, for the DNS entries in the zone that you want toupdate.

A lower value means more hits on your DNS server, while a highervalue means more time for your clients to learn about changes.

The TTL may take the value 0, which indicates never caching therecord, to a maximum of 2,147,483,647, which is over 68 years! The

current best practice recommendation (RFC 1912) proposes a valuegreater than one day, and on RRs that do not change often, considermulti-week values.


Default: 86400

Example: 3600

TTL


120





The list of primary master name servers in the domain.

This attribute is optional since thefirst name server is retrievedfromthe zone ’ s SOA (Start of Authority) record.

If the primary master name server is a stealth server, define thisattribute. A stealth server is a name server that is authoritative for azone, but does not appear in that zone ’ s SOA record. It is hidden toprevent direct attacks from the Internet.


Example: { "10.190.112.23" }

StealthMasters

Required when you configure DNS for secure updates. Specifies theabsolute path to the file containing the private TSIG (TransactionSignature) key. This attribute should be configured only when theDNS server configured is a Unix based DNS server.


Example:

/var/tsig/example.com.+157+00000.private

TSIGKeyFile

Use the CreatePTR attribute to direct the online agent functions to

create PTR records for each RR of type A or AAAA. You must set thevalue of this attribute to true (1) to create the records. Before you canuse this attribute, make sure that the same master or stealth serversserve the forward (A or AAAA) and reverse zones.


Default: 0

Example: 1

CreatePTR

Use the OffDelRR attribute to direct the offline and clean agentfunctions to remove all records that the ResRecord key defines. Youmust set thevalue of this attributeto 1 (true) to have theagent removeall the records.


Default: 0

Example: 1

OffDelRR






Use the UseGSSAPI attribute if the DNS server that you haveconfigured is a Windows DNS server and only if it accepts securedynamic updates.

Note: Do not set this attribute if the Windows DNS server acceptsnon-secure updates.

If this attribute is set to 1, the agent uses the -g option with thensupdate command.

See “ Agent notes for DNS agent ” on page 123. for more informationon requirements to use the DNS agent with the secure Windows DNS

server.


Default: 0

Example: 1

UseGSSAPI

This attribute represents the time interval in seconds after which theDNS agent attempts to refresh theresource records (RRs) on the DNSservers. The default value of zero indicates that the DNS agent doesnotattempt to refresh therecordson theDNS servers. TheDNS agentwrites the warning message to the logs if it is not able to refresh the

DNS records.Note: The refresh request is sent in the next monitor cycle after theRefreshInterval period is reached.

If the DNS agent isunable to refresh the DNS records, and the recordsare removed as a result of a scavenging operation or by the DNSadministrator, the DNS resource will fault.


Default: 0

Example: 3600

RefreshInterval


122





Use this attribute to direct the online agent function to clean up allthe existing DNS records for the configured keys before adding newrecords. The default value (0) disables this behavior.

Note: If multiple DNS resources are configured with the same keyvalue in their ResRecordattribute, then do notset this attributevalueto 1.


Default: 0

Example: 1

CleanRRKeys

Resource type definition for DNS agentThe resource definition for this agent on Linux follows:

type DNS (static keylist SupportedActions = { "dig.vfd","master.vfd", "keyfile.vfd" }static str ArgList[] = { Domain, TTL, TSIGKeyFile,StealthMasters, ResRecord, CreatePTR, OffDelRR,

UseGSSAPI, RefreshInterval, CleanRRKeys }str Domainint TTL = 86400str TSIGKeyFilestr StealthMasters[]str ResRecord{}boolean CreatePTR = 0boolean OffDelRR = 0boolean UseGSSAPI = 0int RefreshIntervalboolean CleanRRKeys = 0

)

Agent notes for DNS agentThe DNS agent has the following notes:

■ About using the VCS DNS agent on UNIX with a secure Windows DNS server

■ High availability fire drill for DNS agent




■ Monitor scenarios for DNS agent

■ Sample Web server configuration for DNS agent

■

Secure DNS update for BIND 9 for DNS agent■ Setting up secure updates using TSIG keys for BIND 9 for DNS agent

About using the VCS DNS agent onUNIXwitha secure WindowsDNS serverThis section describes the requirements for using the DNS agent with a secureWindows DNS server. Note that there are no special requirements for sendingnon-secure updates to a Windows DNS server.

Software requirement for DNS agent

For the secure updates on Windows DNS server to work, the VCS DNS agent onUNIX requires BIND version 9.7.2-P3 or later installed on all cluster nodes.

Configuration requirement for DNS agent

The VCS DNS agent on UNIX requires setting up Kerberos authentication withtheWindows DNSserver andconfiguring thedomain andDNS server informationin /etc/resolv.conf at the client node.

To set up the Kerberos authentication from the UNIX host to the Windows DNSserver, configure the Kerberos configuration file (/etc/krb5.conf or /etc/

krb/krb5.conf) to use the Windows DNS server as Key Distribution Centre (KDC).A sample Kerberos configuration file with domain privdns.sym and DNS servermaster.privdns.sym is as follows:

[libdefaults]default_realm = PRIVDNS.SYMdns_lookup_realm = truedns_lookup_kdc = truedefault_tkt_enctypes = des-cbc-md5default_tgs_enctypes = des-cbc-md5ticket_lifetime = 24hrenew_lifetime = 7dforwardable = trueallow_weak_crypto = true[realms]PRIVDNS.SYM = {kdc = master.privdns.sym:88kpasswd_server = master.privdns.sym:464admin_server = master.privdns.sym


124



}[domain_realm].privdns.sym = PRIVDNS.SYM

privdns.sym = PRIVDNS.SYM

Note: Symantec does not support KDC and Domain Controller/DNS located ondifferent servers.

Authenticate all the nodes on the cluster (on which the DNS agent is configuredto run) with the Active directory. Use kinit on your user account and use klist toverify that you have a ticket to the configured realm principal. Refer to the manpage of kinit for more information on obtaining Kerberos ticket granting ticketsfrom KDC.

Note: The DNS agent requires a node to be authenticated with Kerberos all thetime. Renew the obtained tickets periodically if your authentication methodrequires you to do so.

A sample run of kinit and klist for the above configuration with user vcsdns willlook as follows:

# kinit vcsdnsPassword for [email protected]:

# klistTicket cache: FILE:/tmp/krb5cc_0Default principal: [email protected] starting Expires Service principal12/14/09 16:17:37 12/15/09 02:19:09 krbtgt/[email protected] until 12/21/09 16:17:37

If the environment variable KRB5CCNAME is set to some non-default location(default is /tmp ), then VCS will not inherit it by default and will look for theKerberos tickets in default location /tmp .

To resolve this issue,un-set theenvironment variable KRB5CCNAME andrun thekinit command again. This will update the Kerberos tickets in default location(/tmp ). Else, for a customized location (for example, /cache/krb_ticket ) forKerberos tickets, add an entry in opt/VRTSvcs/bin/vcsenv file on each clusternode before VCS starts:

KRB5CCNAME="FILE:/cache/krb_ticket"

export KRB5CCNAME




Update /etc/resolv.conf onyourclient node toadd information for the WindowsDNS server and the configured domain.

High availability fire drill for DNS agentThehigh availability firedrill detects discrepancies between theVCSconfigurationand the underlying infrastructure on a node; discrepancies that might prevent aservice group from going online on a specific node.

For DNS resources, the high availability drill tests the following conditions:

■ Checks if the key file as specified by the TSIGKeyFile attribute is availableeither locally or on shared storage.

■ Checks if the dig and nsupdate binaries are available on the cluster node andare executable on that node.

■ Checks if the stealth masters can respond to the SOA query made from thecluster node so as to ensure that there is no network issue that would prohibitthe DNS update and query requests from reaching the stealth master server.


Monitor scenarios for DNS agentDepending on the existence of the Online lock file and the defined ResourceRecords (RR), you get different status messages from the Monitor function.

Table 3-11 summarizes the monitor scenarios for the Online lock files.

Table 3-11 Monitor scenarios for the Online lock file

Monitor returnsExpected RR mappingOnline lock fileexists

OFFLINEN/ANO

OFFLINENOYES

ONLINEYESYES

Sample Web server configuration for DNS agentTake the example of a Web server. A browser requests the URLhttp://www.example.com that maps to the canonicalnameserver1.example.com.The browser retrieves the IP address for the web server by querying a domainname server. If the web server fails over from server one to server two(server2.example.com), the domain name servers need a new canonical name


126



mapping forwww.example.com. After the failover, theDNSresourceupdates thismapping of www.example.com to point to canonical name server2.example.com

Note: In this configuration, the Domain attribute shouldbe configured with value"example.com"

Secure DNS update for BIND 9 for DNS agentThe DNS agent expects that the zone ’ s allow-update field contains the IP addressfor the hosts that can dynamically update the DNS records. This functionality isdefault for the DNS agent. Since a competent black hat can, however, spoof IPaddresses, consider TSIG as an alternative.

TSIG (Transaction Signature) as specified in RFC 2845 is a shared key messageauthentication mechanism that is available in BIND DNS. A TSIG key providesthe means to authenticate and verify the validity of exchanged DNS data. It usesa shared secret key between a resolver and either one or two servers to providesecurity.

Setting up secure updates using TSIG keys for BIND 9 for DNSagentIn the following example, the domain is example.com.

To use secure updates using TSIG keys, perform the following steps at the DNS

server:1 Run the dnssec-keygen command with the HMAC-MD5 option to generate

a pair of files that contain the TSIG key:

# dnssec-keygen -a HMAC-MD5 -b 128 -n HOST example.com.

2 Open theexample.com.+157+00000.key file. After you run the cat command,the contents of the file resembles:

# cat example.com.+157+00000.keyexample.com. IN KEY 512 3 157 +Cdjlkef9ZTSeixERZ433Q==

3 Copy the shared secret (the TSIG key), which looks like:

+Cdjlkef9ZTSeixERZ433Q==




4 Configure the DNS server to only allow TSIG updates using the generatedkey. Open the named.conf file and add these lines.

key example.com. {algorithm hmac-md5;secret "+Cdjlkef9ZTSeixERZ433Q==";

};

Where +Cdjlkef9ZTSeixERZ433Q== is the key.

5 In the named.conf file, edit the appropriate zone section and add the allow-updates sub-statement to reference the key:

allow-update { key example.com. ; } ;

6 Save and restart the named process.7 Place the files containing the keys on each of the nodes that are listed in your

group ’ s SystemList. The DNS agent uses this key to update the name server.

Copy both the private and public key files on to the node. A good location isin the /var/tsig/ directory.

8 Set the TSIGKeyFile attribute for the DNS resource to specify the filecontaining the private key.

DNS www (Domain = "example.com"ResRecord = {www = north}TSIGKeyFile = "/var/tsig/example.com.+157+00000.private")

Sample configurations for DNS agentThis section contains sample configurations for this agent.

Basic IPv6 configuration for DNS agentThis sample configuration provides basic configuration for IPv6 support. In thefollowing configuration, nic_value represents thebase NICvalue for the platform

For example: eth0

group ipv6_group_dns (SystemList = { sysA = 0, sysB = 1 })


128



DNS ipv6group_dns_res (Critical = 0Domain = "example.com"

TSIGKeyFile = "/var/tsig/Kipv6.vcscd.net.+157+18435.private"StealthMasters = { "2001:db8:c18:2:69c4:3251:bac1:6cbe" }ResRecord = {

vcssysCv6 = "2001:db8:c18:2:214:4fff:fe96:8833",sysC = vcssysCv6 }

)

IP ipv6group_ip_res (Device @sysA = nic_valueDevice @sysB = nic_valueAddress = "2001:db8:c18:2:214:4fff:fe96:8833"

PrefixLen = 64)

NIC ipv6group_nic_res (Device @sysA = nic_valueDevice @sysB = nic_valueNetworkHosts = { "2001:db8:c18:2:214:4fff:fea2:fd50" }

)

ipv6group_dns_res requires ipv6group_ip_resipv6group_ip_res requires ipv6group_nic_res

IPv6 CNAME sample configuration for DNS agentThe following sample configuration uses CNAME values.

group cname_group (SystemList = { sysA = 0, sysB = 1 })

DNS cname_group_dns_res (Domain = "example.com"StealthMasters = { "3ffe:556::1000:5761" }ResRecord @sysA = { www = server1 }ResRecord @sysB = { www = server2 }OffDelRR = 1

)




IPv4 A sample configuration for DNS agentThe following sample configuration uses A values.

group forwardv4_group (SystemList = { sysA = 0, sysB = 1 })

DNS forward_group_v4_resource (Domain = "example.com"StealthMasters = { "3ffe:556::1000:5761" }

ResRecord @sysA = { www = "10.200.56.240" }ResRecord @sysB = { www = "10.200.56.244" }OffDelRR = 1)

Debug log levels for DNS agentThe DNS agent uses the following debug log levels:



130



File share agents


■ About the file service agents■ NFS agent

■ NFSRestart agent

■ Share agent

■ About the Samba agents

■ SambaServer agent

■ SambaShare agent

■ NetBios agent

About the file service agentsUse the file service agents to provide high availability for file share resources.

NFS agentStarts and monitors the nfsd, mountd, statd, and lockd daemons required by allexported NFS file systems. Configure the NFS resource in a separate parallelservice group with the AutoStart attribute set to 1.

You should configure only a single NFS resource in a service group on a node. If you havemore than one service group that uses theNFS resource, the other servicegroups must use a Proxy resource. The Proxy resource can point to the NFSresource in the first group. Duplicate NFS resources will cause a problem whentheNFSresources arebroughtonline concurrently — only theNFSresource started

4Chapter



first will be successfully brought online, while the rest of the NFS resources mayreport online failure.

For important information about this agent,

See “ Notes for NFS agent ” on page 134.

Dependencies for NFS agentFor more information regarding NFS resource dependencies, refer to the VeritasCluster Server Administrator ’ s Guide .

Figure 4-1 Sample service group that includes an NFS resource

Share

NFSRestart

IP LockMount

DiskGroup

MountNFS/Proxy

NFSRestart

Agent functions for NFS agent

Starts NFSserviceson thesystem.For NFSv3, theservicesalsoincludeNSM (Network Status Monitor) and NLM (Network Lock Manager)service. If NFSv4 is enabled, it also starts NFSv4 name mappingdaemon

Online

Monitors theNFSservices running onthesystem.For NFSv3, it checksfor version 2 and version 3 of NFS service along with other NFSservices. For NFSv4, it also checks the availability of NFSv4 namemapping daemon.

Monitor

Stops and restarts NFS services for all kernels. It also stops andrestarts NSM and NLM services running on the system.

Clean

File share agentsNFS agent

132



State definitions for NFS agent

Indicates that the NFS daemons are running in accordance with the

supported protocols and versions.

ONLINE

Indicates that the NFS daemons are not running in accordance withthe supported protocols and versions.

OFFLINE

Indicates that the NFS daemons are not running in accordance withthe supported protocols and versions.

FAULTED

Unable to determine the status of the NFS daemons.UNKNOWN

Attributes for NFS agentTable 4-1 Optional attributes for Linux

DescriptionOptionalattributes

Requiredwhen thevalueof theNFSRestart attributeis 1. GracePeriodspecifies the amount of time that lock recovery is allowed by the NFSserver after its reboot.


Default: 90

GracePeriod

The NFS and the NFSRestart agents require a synchronizationmechanism when the group to which they belong is in transition, forexample going online or coming offline. A file serves as thissynchronizationmechanism.The LockFileTimeoutattributespecifiesthe maximum time that the synchronization file exists.


Default: 180

LockFileTimeout

Specifies the number of concurrent NFS requests that the server canhandle.


Default: 8

Example: 16

Nproc




Table 4-1 Optional attributes for Linux (continued)

DescriptionOptionalattributes

Specifies whether to start the NFS security daemon rpc.svcgssd ornot.

You must configure the type of security that NFS supports, forexample: Kerberos.


Default: 0

NFSSecurity

Specifies whether to start the NFSv4 daemon rpc.idmapd or not andwhether to monitor nfsd version 4.


Default: 0

NFSv4Support

Resource type definition for NFS agenttype NFS (

static int RestartLimit = 1static str Operations = OnOnlystatic str ArgList[] = { Nproc, GracePeriod, NFSSecurity,NFSv4Support, LockFileTimeout }int Nproc = 8int GracePeriod = 90boolean NFSSecurity = 0boolean NFSv4Support = 0int LockFileTimeout = 180

)

Notes for NFS agentThe NFS agent has the following notes:

■ Prerequisites for NFS lock recovery

■ Using NFSv4

Prerequisites for NFS lock recoveryIf you plan on using lock recovery on a Linux system, store locking informationon sharedstorage so that it is accessible to the systemwhere NFS fails over. Usingthis information, NFS carries out lock recovery.


134



For more information, refer to the NFSRestart agent.

Using NFSv4The NFS agent provides NFSv4 support to export shares using the attributeNFSv4Support. Only oneof theShare resources that depends on theNFSresourceneeds a value of fsid=0 in its Options attribute. The shared directory that has thefsid=0 option becomes the root of all exports. The client needs to mount only thisroot file system instead of mounting all shares individually.

The syntax is:

mount -t nfs4 <server>:/ <mountpoint>

Always use a slash (/) to end the path after the colon (:).

All the file systems, other than the root file system, needs to have the nohideoption set in Options attribute of share resources. Set the nohide option so thatauthentic clients can seamlessly move through the tree of exported file systems just by mounting the root file system.

To enable NFSv4 support on your node, you must have the rpc_pipefs (pipe filesystem) mounted on the node. At boot time, rpc_pipefs is mounted on every Linuxnode. In situations where it is notmounted, mount rpc_pipefs on thecluster node.

To mount rpc_pipefs

At the prompt on the node, enter the following:

# mount -t rpc_pipefs rpc_pipefs /var/lib/nfs/rpc_pipefs

Sample configurations for NFS agentOn each node in your cluster, you can find sample NFS, NFSRestart, and Shareconfigurations in /etc/VRTSvcs/conf/sample_nfs/.

For more information regarding agent configuration, refer to the Veritas Cluster Server Administrator ’ s Guide .

Debug log levels for NFS agentThe NFS agent uses the following debug log levels:





NFSRestart agentThe NFSRestart agent provides the following functionalities:

Manages NFS lock recovery service by recovering the NFS record locks aftersudden server crash.

Prevents potential NFS ACK storms by terminating NFS server services beforeoffline of NFS VIP to close all TCP connections with the NFS client.

If you have configured the NFSRestart agent for lock recovery, the NFSRestartagent starts the smsyncd daemon. The daemon copies the NFS locks from thelocal directory /var/lib/nfs to shared storage. The agent ’ s online function copiesthe locks from shared storage to local directory /var/lib/nfs.

For important information about this agent, refer to Notes for NFSRestart agent

Dependencies for NFSRestart agentFor more information regarding NFSRestart resource dependencies, refer to theVeritas Cluster Server Administrator ’ s Guide .

You must use two NFSRestart resources in a service group. Both the NFSRestartresources provide combined protection from potential corruption of NFS locksand potential NFS ACK storms. The lower NFSRestart resource must have itsLower attribute set to 1. The upper NFSRestart resource should be at the top of the resource dependency tree and the lower NFSRestart resource shouldbe below

the Share resource in the resource dependency tree. The NFSRestart resourcesand the Share resources must be inside the same service group.

File share agentsNFSRestart agent

136



Figure 4-2 Sample service group that includes an NFSRestart resource

Share

NFSRestart

IP LockMount

DiskGroup

MountNFS/Proxy

NFSRestart

Agent functions for NFSRestart agentThe agent functions for this agent follow:

For the lower NFSRestart resource:

■ If the value of the NFSLockFailover attribute is 1, the agent

terminates statd and lockd.For the upper NFSRestart resource:

■ If the value of the NFSLockFailover attribute is 1, the agent copiestheNFSrecord locks fromshared storage to /var/lib/nfsdirectory.

■ Starts the statd and lockd daemons.■ Starts the smsyncd daemon to copy the contents of /var/lib/nfs

directory to the shared storage (LocksPathName) at regular twosecond intervals.

Online


■ The monitor agent function does nothing.For the upper NFSRestart resource:

■ If the value of the NFSLockFailover attribute is 1, the agentmonitors smsyncd daemon. It restarts the smsyncd daemon if itis not running.

Monitor





■ Restarts all the NFS daemons that the upper NFSRestart resourcestopped previously.

For the upper NFSRestart resource:

■ Terminates the statd and lockd daemons to clear the lock state.■ Terminates the nfsd and mountd daemons to close the TCP/IP

connections.■ Terminates the smsyncd daemon if the daemon is running.

Offline


■ Restarts all the NFS daemons that the upper NFSRestart resourcestopped previously.

For the upper NFSRestart resource:

■ Terminates the statd and lockd daemons to clear the lock state.■ Terminates the nfsd and mountd daemons to close the TCP/IP

connections.■ Terminates the smsyncd daemon if the daemon is running.

Clean

■ nfsconf.vfdChecks the runlevel information of the system service nfslock toconfirm that the lock daemons do not come online automaticallyafter reboot.

■ lockdir.vfd

Verifies that the NFS lock directory (which is specified by theLocksPathName attribute of NFSRestart) is on shared storage.

Action

State definitions

Indicates that the daemons are running properly.ONLINE

Indicates that one or more daemons are not running.OFFLINE

Indicates the inability to determine the agent ’ s status.UNKNOWN


138



Attributes for NFSRestart agentTable 4-2 Required attributes


Name of the NFS resource on the system. This attribute is required if the value of the NFSLockFailover attribute is 1.


Example: "nfsres1"

NFSRes



The path name of the directory to store the NFS lock information.This attribute is required when the value of the NFSLockFailoverattribute is 1. The path that you specify for the LocksPathNameattribute should be on shared storage. This is to ensure that it isaccessible to allthe systemswhere theNFSRestart resource failsover.


Example: "/share1x"

LocksPathName

NFS Lock recovery is done for all the Share resources that areconfigured in the group of this resource.


Default: 0

NFSLockFailover

Defines the position of NFSRestart resource in theservice group. TheNFSRestart resource below the Share resource needs a value of 1.

The NFSRestart resource on the top of the resource dependency treehas a Lower attribute value of 0.


Default: 0

Lower

Resource type definition for NFSRestart agenttype NFSRestart (static keylist SupportedActions = { "lockdir.vfd", "nfsconf.vfd"}




static str ArgList[] = { "NFSRes:Nproc", "NFSRes:GracePeriod","NFSRes:NFSv4Support", NFSLockFailover, LocksPathName, Lower,State }

str NFSResint Lower = 0str LocksPathNameboolean NFSLockFailover = 0)

Notes for NFSRestart agentThe NFSRestart agent has the following notes:

■ About high availability fire drill

■ Providing a fully qualified host name

About high availability fire drillThehigh availability firedrill detects discrepancies between theVCSconfigurationand the underlying infrastructure on a node; discrepancies that might prevent aservice group from going online on a specific node.

For NFSRestart resources, the high availability drill performs the following, it:

■ Checks the NFS configuration file to confirm that the NFS server does notcome online automatically after reboot.

■ Verifies that theNFSlock directory (which is specified by the LocksPathNameattribute of NFSRestart) is on shared storage.


Providing a fully qualified host nameYou must provide a fully qualified host name, forexample, nfsserver.example.edu,for the NFS server while mounting the file system on the NFS client. If you do notuse a fully qualified host name, or if you use a virtual IP address (10.122.12.25)or partial host name (nfsserver), NFS lock recovery may fail.

If you want to use the virtual IPaddress ora partial host name, makethe followingchanges to the service database (hosts) and the nsswitch.conf files:

/etc/hosts


140



To use the virtual IP address and partial host name for the NFS server, you needto add an entry to the /etc/hosts file. The virtual IP address and the partial hostname should resolve to the fully qualified host name.

/etc/nsswitch.conf

You should also modify the hosts entry in this file so that upon resolving a namelocally, the host does not first contact NIS/DNS, but instead immediately returnsa successful status. Changing the nsswitch.conf file might affect other servicesrunning on the system.

For example:

hosts: files [SUCCESS=return] dns nis

You have to make sure that the NFS client stores the same information for theNFS server as the client uses while mounting the file system. For example, if theNFS client mounts the file system using fully qualified domain names for the NFSserver, then the /var/statmon/smdirectory on the NFSclient should also containa fully qualified domain name of the NFS server after the acquisition of locks.Otherwise you need to stop and start the status daemon and lock daemon to clearthe lock cache of the NFS client.

A time period exists where the virtual IP address is online but locking servicesare not registered on the server. Any NFS client trying to acquire a lock in thisinterval would fail and get ENOLCK error.

Every two seconds, the smsyncd daemon copies the list of clients that hold thelocks on the shared filesystem in theservicegroup. If theservicegroupfails beforesmsyncd has a chance to copy the client list, the clients may not get a notificationonce the service group is brought up. This causes NFS lock recovery failure.

Sample configurations for NFSRestart agentOn each node in your cluster, you can find sample NFS, NFSRestart, and Shareconfigurations in /etc/VRTSvcs/conf/sample_nfs/.

For more information regarding agent configuration, refer to the Veritas Cluster

Server Administrator ’

s Guide .

Basic agent configurationsFor NFS lock recovery:

NFSRestart nfsrestart (NFSRes = nfsresLocksPathName="/shared_mnt/lockinfo"




NFSLockFailover = 1Lower = 0)

NFSRestart nfsrestart_L (NFSRes = nfsresLocksPathName="/shared_mnt/lockinfo"NFSLockFailover = 1Lower = 1)

For no NFS lock recovery:

NFSRestart nfsrestart (NFSRes = nfsres)NFSRestart nfsrestart_L (NFSRes = nfsresLower = 1)

Debug log levels for NFSRestart agentThe NFSRestart agent uses the following debug log levels:


Share agentShares, unshares, and monitors a single local resource for exporting an NFS filesystem to be mounted by remote systems.

Before you use this agent, verify that the files and directories to be exported areon shared disks.

For important information on this agent, refer to:

Notes for Share agent

Dependencies for Share agentFormore information regarding Share resourcedependencies, refer to the VeritasCluster Server Administrator ’ s Guide .

Share resources dependon NFS. Inan NFS service group, the IPfamily of resourcesdepends on Share resources.

File share agentsShare agent

142





The Client attribute can be assigned one FQDN host name or IPaddress, whereas multiple FQDN host names, or IP addresses can beassignedto the OtherClients field.Formore information onspecifyingIP address, refer to About enteringan IP address in theClientattribute .

A combination of 'Client' and 'OtherClients' can be used to specify thehost names.

If both of the Client and OtherClients attributes are left unspecified,the PathName is exported to the world (*).


OtherClients

Resource type definition for Share agenttype Share (static keylist SupportedActions = { "direxists.vfd" }static str ArgList[] = { PathName, Client, OtherClients,Options, "NFSRes:State" }str PathNamestr Clientstr OtherClients[]

str Optionsstr NFSRes)

Notes for Share agentThe following section contains notes on the Share agent.

■ Support for spaces in directory names


■ About entering an IP address in the Client attribute

Support for spaces in directory namesThe Share agent supports directory names with spaces. The space can be leading,trailing, or in the middle of the name. If the directory name has a trailing space,provide an extra "/" at the end of the PathName attribute of a Share resource.Note that the agent does not support spaces created using the TAB key.

File share agentsShare agent



High availability fire drillThehigh availability firedrill detects discrepancies between theVCSconfigurationand the underlying infrastructure on a node. These discrepancies might preventa service group from going online on a specific node. For Share resources, thehigh availability fire drill checks if the path exists.


About entering an IP address in the Client attributeYou can enter an IP address as a part of the Client or OtherClients attribute. If you specify an IPaddress, use the form a.b.c.d/32 . Using this form of IPaddressprevents the monitor entry point from reporting a Share resource as offline even

if the sharetab entry is present. This occurs on some SLES systems where theexportfs command tries to resolve the client IP address to the host name andrewrite the sharetab entry with the host name. If you specify the IP address asa.b.c.d/32 , exportfs does not recognize this as an IP address and does notattempt to resolve the address to a host name.

Sample configurations for Share agentOn each node in your cluster, you can find sample NFS, NFSRestart, and Shareconfigurations in /etc/VRTSvcs/conf/sample_nfs/.

For more information regarding agent configuration, refer to the Veritas Cluster Server Administrator ’ s Guide .

Debug log levels for Share agentThe Share agent uses the following debug log levels:


About the Samba agentsSamba is a suite of programs that allows a system running a UNIX or UNIX-likeoperating systemto provide services using theMicrosoft networkprotocol. Sambasupports the following services:

■ Filespace

■ Printer

■ WINS

File share agentsAbout the Samba agents

146



■ Domain Master

Configure these services in the Samba configuration file (smb.conf). Samba usestwo processes: smbd and nmbd to provide these services.

VCS provides Samba failover using three agents: SambaServer, NetBios, andSambaShare.

The Samba agents■ The NetBios agent

■ The SambaServer agent

■ The SambaShare agent

Before using the Samba agents■ Verify that smbd and nmbd always run as daemons. Verify that they cannot

be started using the meta-daemon inetd.

■ Verify that thesmbd andnmbd daemons arein thepath environment variable.The default path of the smbd and nmbd daemons is: /usr/sbinFor more information on configuring these paths, refer to the description of the SambaTopDir attribute.

■ Verify that Samba is configured properly and that the Samba configuration

file is identical on all cluster systems. The user can replicate the file or storeit on a shared disk accessible from all cluster systems.

■ If configuring Samba as a WINS server or Domain Master, verify that theSamba lock directory is on the shared disk. This ensures that the WINS serverdatabase and Domain Master are created on the shared disk.

Supported versions for Samba agentsVCS Samba suite of agents support Samba version 3.0 and above. Please checkyour samba version using the following command:

# smbd -V

Notes for configuring the Samba agentsThe following notes describe configuration considerations for the Samba agents.

File share agentsAbout the Samba agents



To check for the default location of the Samba pidfile

Enter the following command:

# smbd -b | grep PIDDIR

SambaServer agentTheSambaServer agent starts, stops, and monitors thesmbd processas a daemon.Only one resource of this type is permitted. You can use the agent to make a smbddaemon highly available.

The smbd daemon provides Samba share services. The agent verifies that Sambais running by reading the pid of smbd daemon. The agent can perform in-depthmonitoring by establishing a socket connection to Samba at ports where thedaemon is listening and sending it a NetBIOS session request.

Dependencies for SambaServer agentNo dependencies exist for the SambaServer resource.

Figure 4-4 Sample service group that includes a SambaServer resource

SambaShare

NetBIOS SambaServer

IP

Agent functions for SambaServer agent

Starts the smbd daemon at specified or default ports.Online

Stops the smbd daemon.Offline

Verifies that the smbd daemon is running by reading its pid file. Doesindepth monitoringperiodically, if configured, by establishing a socketconnection to Samba and sending it a NetBIOS session request.

Monitor

Stops the smbd daemon forcefully if required.Clean

File share agentsSambaServer agent



State definitions for SambaServer agent

Indicates that the smbd daemon is running. If in-depth monitoring is

configured, it indicates that a positive session response packet wasreceived through a socket connection to the Samba server.

ONLINE

Indicates that smbd is not running. If in-depth monitoring is enabled,it indicates that the agent couldnot establish a socket connection withthe server, or that it received an incorrect response packet header, orthe session response packet connection timed out.

OFFLINE

Indicatesthat theagentcouldnotdetermine thestateof theresource.UNKNOWN

Indicates that the smbd daemon has stopped unexpectedly or is notresponding (if in-depthmonitoringis enabled)outsideof VCScontrol.

FAULTED

Attributes for SambaServer agentTable 4-6 Required attributes


Complete path of the configuration file that Samba uses.


Example: "/etc/samba/smb.conf"

ConfFile

Lock directory of Samba. Samba stores the files smbd.pid, nmbd.pid,wins.dat (WINS database), andbrowse.dat (master browser database)in this directory.


Example:"/var/run"

LockDir



Number of monitor cycles after which the in-depth monitoring is performed. Forexample, thevalue5 indicates that theagent monitors theresource in-depth everyfive monitor cycles. The value 0 indicates that theagent will notperform in-depthmonitoring for the resource.


Default: 5

IndepthMonitorCyclePeriod


150





Ports where Samba accepts connections.To run Samba over NBT (NetBios over TCP/IP), set this attribute to 139. To runSamba directly over TCP/IP, set this attribute to 445.

Type and dimension: integer-vector

Default: 139, 445

Ports

Number of seconds the agent waits to receive the session response packet aftersending the session request packet. For example, the value 5 indicates that theagentwaits forfive seconds beforereceivingthe session responsepacket.Configurethis attribute if in-depth monitoring is enabled.


Default: 10

ResponseTimeout

The absolute path to the Samba daemon pid file. This file contains the process IDof the monitored smbd process.

Configure this attribute if you are using a non-standard configuration file nameor path. If this attribute is not configured for non-standard configuration filenames, the agent checks the smbd- ConfFile .pid file for monitoring the resource.


Example: "/var/run/smbd.pid"

PidFile

Parent path of Samba daemon and binaries.


Example: "/usr"

SambaServer agent uses SambaTopDir attribute value in an open entry point todetermine the complete path of samba executables. If this attribute is configuredafter theresource is enabled, pleasedisable andenable theresourceagain to bringthis into effect as follows:

# hares -modify <res> Enabled 0

# hares -modify <res> Enabled 1

SambaTopDir






TheIP addresswhere theSambadaemon (smbd) listensforconnections.ConfiguretheSocketAddressattribute if youareconfiguring multiple SambaServerresourceson a node.

Note: Only IPv4 addresses are supported.

Type and Dimension: string-scalar

Example: “ 10.128.10.14 ”

SocketAddress

Resource type definitions for SambaServer agent

type SambaServer (static str ArgList[] = { ConfFile, LockDir, Ports,IndepthMonitorCyclePeriod, ResponseTimeout, SambaTopDir,PidFile, SocketAddress}str ConfFilestr LockDirint Ports[] = { 139, 445 }int IndepthMonitorCyclePeriod = 5int ResponseTimeout = 10str SambaTopDirstr PidFilestr SocketAddress)

Sample configurations for SambaServer agentThe sample configurations for this agent follow:

SambaServer samba_server (ConfFile = "/etc/samba/smb.conf"LockDir = "/usr/lock/samba"

IndepthMonitorCyclePeriod = 3ResponseTimeout = 15)

Debug log levels for SambaServer agentThe SambaServer agent uses the following debug log levels:



152



SambaShare agentThe SambaShare agent adds, removes, and monitors a share by modifying the

specified Samba configuration file. You can use the agent to make a Samba Sharehighly available.

Each filespace service provided by Samba is a shared resource and is defined asa section in the Samba configuration file. The section name is the name of theshared resource and the section parameters define the share attributes.

Dependencies for SambaShare agentSambaShareresources dependon theSambaServer,NetBiosandMount resources.

Figure 4-5 Sample service group for a SambaShare resource

SambaShare

NetBIOS SambaServer

IP

Agent functions for SambaShare agent

Edits the samba configuration file and adds the shares.Online

Removes the shares from the configuration file.Offline

Issues the command smbclient to check if thespecified shares exist.Monitor

Terminates all ongoing connectionswith the particular samba share,removes its entry from the samba configuration file and reloads theconfiguration.

Clean

State definitions for SambaShare agent

Indicates that the share is available.ONLINE

Indicates that the share is not available.OFFLINE

Indicatesthat thesharehas becomeunavailableoutside ofVCScontrol.FAULTED


File share agentsSambaShare agent



Attributes for SambaShare agentTable 4-8 Required attributes


Name of the SambaServer resource.


Example: "smb_res1"

SambaServerRes

Name of the share resource as exported by samba.

Note: This name can be different from the SambaShare resourcename.


Example: "share1"

ShareName

List of parameters for the share attributes. These parameters arespecifiedas name=valuepairs,with each pair separatedby a semicolon(;).


Example: "path=/shared; public=yes; writable=yes"

ShareOptions

Resource type definition for SambaShare agenttype SambaShare (static str ArgList[] = { "SambaServerRes:ConfFile","SambaServerRes:LockDir", ShareName, ShareOptions,"SambaServerRes:Ports", SambaServerRes,"SambaServerRes:SambaTopDir", "SambaServerRes:PidFile","SambaServerRes:SocketAddress" }str SambaServerResstr ShareNamestr ShareOptions)

Sample configuration for SambaShare agentSambaShare Samba_SambaShare3 (SambaServerRes = Samba_SambaServerShareName = smbshare3

File share agentsSambaShare agent

154



ShareOptions = "path=/smbshare3; public=yes; writable=yes")

Debug log levels for SambaShare agentThe SambaShare agent uses the following debug log levels:

DBG_1, DBG_3, DBG_5

NetBios agentTheNetBiosagent starts, stops,andmonitors thenmbd daemon. Only oneresourceof this type is permitted. You can use the agent to make the nmbd daemon highlyavailable.

The agent sets, monitors, and resets the names and network interfaces by whichthe Samba server is known. The agent also sets, monitors and resets Samba to actas a WINS server or domain master or both.

Note: The nmbd broadcasts the NetBIOS name, or the name by which the Sambaserver is known in the network.

Dependencies for NetBios agent

The NetBios resource depends on the IP or the IPMultiNIC resource if the virtualIPaddress configuredin theIP/IPMultiNIC resource is being used in theInterfacesattribute of the NetBios resource.

Figure 4-6 Sample service group that includes a NetBIOS resource

SambaShare

NetBIOS SambaServer

IP

File share agentsNetBios agent



Agent functions for NetBios agent

Updates theSamba configuration withtheNetBIOSname, allNetBIOS

aliases and network interfaces, WINS support, and domain masteroptions specified in the NetBIOS resource. Starts the nmbd daemon.

Online

Removes the NetBIOS name, all NetBIOS aliases and networkinterfaces, WINS support, anddomain masteroptions specifiedin theNetBIOS resource from theSamba configuration file. Stops the nmbddaemon.

Offline

Verifies that theSamba configuration contains theNetBIOS name, allNetBIOS aliases and network interfaces, WINS support, and domainmaster options specified in the NetBIOS resource. Also verifies thatthe nmbd daemon is running by reading its pid file.

Monitor

Removes the NetBIOS name, all NetBIOS aliases and networkinterfaces, WINS support, anddomain masteroptions specifiedin theNetBIOS resource from theSamba configuration file. Stops the nmbddaemon, forcibly when necessary.

Clean

State definitions for NetBios agent

Indicates that the specified NetBIOS name and aliases are advertisedand that Samba is handling requests for all specified networkinterfaces. Indicates that WINS and Domain support services arerunning, if configured.

ONLINE

Indicates one or more of the following:

■ NetBIOS name is not advertised.■ A NetBIOS alias is not advertised.■ Samba is not handling requests on any of the specified interfaces.■ If WINS support is configured, Samba is not providing WINS

service.■ If domain support is set, Samba is not providing Domain Master

service.

OFFLINE


Indicates that the resource has become offline unexpectedly outsideof VCS control.

FAULTED


156



Attributes for NetBios agentTable 4-9 Required attributes


Name by which the Samba server is known in the network.


Example: "samba_demon"

Note: Samba hasa limitation of 15 characters for NetBios names andaliases.

NetBiosName

Name of the SambaServer resource.


Example: "smb_res1"

SambaServerRes



List of network interfaces on which Samba handles browsing.


Example: "172.29.9.24/16"

Note: If you have configured the SocketAddress attribute value forthecorrespondingSambaServer resource,then youmustalsoconfigurethe same value paired with the appropriate netmask in the list of interfaces.

Interfaces

List of additional names by which the Samba server is known in thenetwork.


Example: { host1_samba, myname }

Note: Samba hasa limitation of 15 characters for NetBios names andaliases.

NetBiosAliases

If set to 1, this flag causes the agent to configure Samba as a WINSserver.


Default: 0

WinsSupport




WinsSupport = 1DomainMaster = 1)

Debug log levels for NetBios agentThe NetBios agent uses the following debug log levels:

DBG_1, DBG_5





160



Service and applicationagents


■ About the services and applications agents

■ Apache HTTP server agent

■ Application agent

■ CoordPoint agent

■ KVMGuest agent

■ Process agent

■ ProcessOnOnly agent

About the services and applications agentsUseservice andapplication agents to provide high availability forapplication andprocess-related resources.

Apache HTTP server agentThe Apache HTTP server agent brings an Apache Server online, takes it offline,and monitors its processes. The Apache HTTP server agent consists of resourcetype declarations and agent scripts. You use the Apache HTTP server agent, inconjunction with other agents, to make an Apache HTTP server highly available.

This agent supports the Apache HTTP server 2.0 and 2.2. It also supports the IBMHTTP Server 1.3, 2.0 and 7.0.0.0.

5Chapter



This agent can detect when an Apache HTTP server is brought down gracefullyby an administrator. When Apache is brought down gracefully, the agent doesnot trigger a resource fault even though Apache is down.

Note: The Apache agent requires an IP resource for operation.

For more information regarding this agent:

See “ Apache HTTP server notes ” on page 168.

DependenciesThis type of resource depends on IP and Mount resources.

Figure 5-1 Sample service group for the Apache HTTP server agent

Apache

IP Mount

NIC Volume

DiskGroup

Agent functions

To start the Apache HTTP server, the agent:

■ Executes the httpdDir/httpd program with the appropriatearguments if the httpdDir program specifies the full path of thedirectory in which the httpd binary file is located.

■ Alternatively, if the httpdDir attribute specifies the full path of the Apache HTTP server binary file, the binary file is executedwith appropriate arguments.

When you specify a file with the EnvFile attribute, the file is sourcedbefore the agent executes the Apache HTTP server commands.

Online

Service and application agentsApache HTTP server agent

162



To stop the Apache HTTP server, the agent:

■ Executes the httpdDir/httpd program with the appropriatearguments, if httpdDir specifies the full path of the directory inwhich the httpd binary file is located.

■ Alternatively, if the httpdDir attribute is used to specify the fullpath of theApache HTTP serverbinary, the binary file is executedwith appropriate arguments.

■ Sends a TERM signal to theHTTP Serverparent process (Apache).

When you specify a file with the EnvFile attribute, the file is sourcedbefore the agent executes the Apache HTTP server commands.

Offline

Monitors the state of the Apache server. First it checks for theprocesses, next it can perform an optional state check.

Monitor

Removes theApacheHTTP serversystemresources thatmight remainafter a serverfault or after anunsuccessfulattempt toonlineor offline.These resources include the parent httpd daemon and its childdaemons.

Clean

checkconffile.vfd

Checks for the existence of the Apache configuration file and theexistence of the directory that contains the httpd binary that is usedduring start up.

Fora local installation, if theconfig file or HttpdDir isnotfound, makesure that it exists on the failover node.

Action

State definitions

Indicates that the Apache server is running.ONLINE

Indicates that the Apache server is not running.

Canalso indicate that theadministrator hasstopped theHTTP servergracefully.Note that theagentuses thePidFile attributeforintentionaloffline detection.

OFFLINE

Indicates that a problem exists with the configuration.UNKNOWN






Full path and file name of the main configuration file for theApache server.


Example: "/apache/server1/conf/httpd.conf"

ConfigFile

Full path of the Apache HTTP server binary file or full path of the directory in which the httpd binary file is located.


Example: "/apache/server1/bin"

httpdDir

This attribute is required when youwant to enable thedetectionof a graceful shutdown outside of VCS control.

See Table 5-2 on page 164.

PidFile



A list of directives that httpd processes after reading the

configuration file.Type and dimension: string-association

Example: DirectiveAfter{} = { KeepAlive=On }

DirectiveAfter

A list of directives that httpd processes before it reads theconfiguration file.


Example: DirectiveBefore{} = { User=nobody, Group=nobody }

DirectiveBefore

Account name the agent uses to execute the httpdprogram. If

you do not specify this value, the agent executes httpd as theroot user.


Example: "apache1"

User


164





If this attribute is set to 1 (true) the online agent function willadd support for SSL, by including the option -DSSL in thestart command.

For example: /usr/sbin/httpd -f path_to_httpd.conf -k start -DSSL

Where path_to_httpd.conf fileis thepathto thehttpd.conf file.

If this attribute is set to 0 (false) the agent excludes the SSLsupport.


Default: 0Example: "1"

EnableSSL

The virtual host name that is assigned to the Apache serverinstance.The host name isused insecond-levelmonitoringforbenchmarking the Apache HTTP server.

You canuse IPv4 or IPv6 addresses for the HostName attribute.

Note: The HostName attributeis only requiredwhenthevalueof SecondLevelMonitor is 1 (true).


Example: "web1.example.com"

HostName

Port number where the Apache HTTP server instance listens.The port number is used in second-level monitoring forbenchmarking the Apache HTTP server. Specify this attributeonly if SecondLevelMonitor is set to 1 (true).


Default: 80

Example: "80"

Port






Full path and file name of the file that is sourced beforeexecuting Apache HTTP server commands. With Apache 2.0,the file ServerRoot /bin/envvars, which is supplied in mostApache 2.0 distributions, is commonly used to set theenvironment before executing httpd. Specifying this attributeis optional. If EnvFile is specified, the shell for user must beBourne, Korn, or C shell.


Example: "/apache/server1/bin/envvars"

EnvFile

The PidFile attribute sets the file to which the server recordsthe process ID of the daemon. The value of PidFile attributemust be the absolute path where the Apache instance recordsthe pid.

This attribute is required when you want the agent to detectthe graceful shutdown of the Apache HTTP server. For theagent to detect the graceful shutdown of the Apache HTTPserver, the value of the IntentionalOffline resource typeattribute must be 1 (true).


Example: /var/run/httpd.pid

PidFile

Full path of the directory in which the Apache HTTP sharedobject files are located. Specifying this attribute is optional. Itis used when the HTTP Server is compiled using theSHARED_CORE rule.If youspecify thisattribute, the directoryis passed to the -R option when executing the httpd program.Refer to the httpd man pages for more information about the-R option.


Example: "/apache/server1/libexec"

SharedObjDir

Enablessecond-levelmonitoring for the resource. Second-levelmonitoring is a deeper, more thorough state check of theApache HTTP server. Valid attribute values are 1 (true) and 0(false).


Default: 0

Example: "1"

SecondLevelMonitor


166





The number of seconds that the monitor agent function waitson the execution of second-level monitor. If the second-levelmonitor program does not return to calling the monitor agentfunction before the SecondLevelTimeout window expires, themonitor agent function no longer blocks on the programsub-process.It does, however, report that theresourceisoffline.The value should be high enough to allow the second levelmonitor enough time to complete. The value should be lessthan the value of the agent's MonitorTimeout.


Default: 30 Table

SecondLevelTimeout

Controls the agent ’ s logging detail for a specific instance of aresource. Values are

■ ERROR: Logs error messages.■ WARN: Logs error and warning messages■ INFO: Logs error, warning, and informational messages.■ TRACE: Logs error, warning, informational, and trace

messages. Trace logging is verbose. Use for initialconfiguration or troubleshooting.


Default: INFO

Example: "TRACE"

ResLogLevel

Table 5-3 Resource type attribute


For information on how to use the IntentionalOffline resource typeattribute, refer to the Veritas Cluster Server Administrator ’ s Guide .

IntentionalOffline

Resource type definition

type Apache (static keylist SupportedActions = { "checkconffile.vfd" }static str ArgList[] = { ResLogLevel, State, IState, httpdDir,SharedObjDir, EnvFile, PidFile, HostName, Port, User,




SecondLevelMonitor, SecondLevelTimeout, ConfigFile, EnableSSL,DirectiveAfter, DirectiveBefore }str ResLogLevel = INFO

str httpdDirstr SharedObjDirstr EnvFilestr PidFilestr HostNameint Port = 80str Userint SecondLevelTimeout = 30str ConfigFile = 0str DirectiveAfter{}str DirectiveBefore{}

boolean EnableSSL

static boolean IntentionalOffline = 0)

Apache HTTP server notesThe Apache Apache HTTP server has the following notes:

■ See “ Tasks to perform before you use the Apache HTTP server agent ”

on page 168.

■ See “ About detecting application failure ” on page 169.■ See “ About bringing an Apache HTTP server online outside of VCS control ”

on page 169.

■ See “ About high Availability fire drill ” on page 170.

Tasks to perform before you use the Apache HTTP server agentBefore you use this agent, perform the following tasks:

■ Install the Apache server on shared or local disks.

■ Ensure that you are able to start the Apache HTTP server outside of VCScontrol, with the specified parameters in the Apache configuration file (forexample: /etc/apache/httpd.conf). For more information on how to start theserver:See “ About bringing an Apache HTTP server online outside of VCS control ”

on page 169.

■ Specify the location of the error log file in the Apache configuration file foryour convenience (for example: ErrorLog /var/apache/logs/error_log).


168



■ Verify that the floating IP has the same subnet as the cluster systems.

■ If you use a port other than the default 80, assign an exclusive port for theApache server.

■ Verify that the Apache server configuration files are identical on all clustersystems.

■ Verify that the Apache server does not autostart on system startup.

■ Verify that inetd does not invoke the Apache server.

■ The service group has disk and networkresources to support theApacheserverresource.

■ Assign virtual host name and port to Apache Server.

About detecting application failureThe agent provides two methods to evaluate the state of an Apache HTTP serverinstance. The first state check is mandatory and the second is optional.

The first check determines the state of the Apache HTTP server. The checkdetermines the state by searching for the existence of the parent httpd daemon.It also searches for at least one child httpd daemon. If the parent process and atleast one child do not exist, VCS reports the resource as offline. If they do exist,and if the agent attribute SecondLevelMonitor is set to true, then the Apacheagent uses the Apache Benchmarking utility "ab" to perform detail monitoring.

If the exit code of the "ab" utility is 0 and if the command output contains"Benchmarking HostName", the agent considers the server online, else the agentconsiders the server offline.

If the binary file ab is not found, Apache agent uses the ab2 binary file for detailmonitoring.

About bringing an Apache HTTP server online outside of VCScontrolWhen you bring an Apache HTTP server online outsideofVCScontrol, first sourceitsenvironment file. Start theserver with the -f option so theserverknows whichinstance to start. You can then specify additional options (such as EnableSSL orSharedObjDir) that you want the server to use at start.




To start an Apache HTTP server outside of VCS control

1 Source the environment file if required.

2 Start the Apache HTTP server. You must use the -f option so that the agentcan distinguish different instances of the server.

httpdDir /httpd -f ConfigFile -k start

Where httpdDir is /apache/v2.2/bin ConfigFile is /apache/v2.2/conf/httpd.conf.When fully formed, the start example looks like:

/apache/v2.2/bin/httpd -f /apache/v2.2/conf/httpd.conf -k start

3 Specify additional options such as EnableSSL or SharedObjDir that you wantto use when you start server. When you add EnableSSL to the command, itresembles:

httpdDir /httpd -f ConfigFile -k start -DSSL

Note: You can specify the full path of a binary file without having httpd aspart of httpdDir attribute.

For example:

/usr/sbin/apache2 -f /etc/httpd/conf/httpd.conf -k start

About high Availability fire drillThehigh availability firedrill detects discrepancies between theVCSconfigurationand the underlying infrastructure on a node. These discrepancies might preventa service group from going online on a specific node.

ForApache resources, when the ApacheHTTP server is installed locally, the highavailability fire drill checks for the validity of these attributes:

■ ConfigFile

■ httpdDir



170



Sample configurations

Running two versions of httpd for LinuxThis example shows how two versions of httpd can run from different locations.In group Apache_1, httpdruns from Port 80, thedefault location.Theconfigurationfile in /usr/local/apache/conf/httpd.conf shouldindicateDocumentRoot, address,port, and other parameters. In group Apache_2, httpd runs from /home/web/apache. The PID file for this is created in /home/web/apache/logs/httpd.pid. The configuration file in /home/web/apache/conf/httpd.conf should define parameters for running thisversion of httpd.

Each Apacheresource requiresan online IP resource. In thisexample, each Apache

resource requires an online mount resource to mount block devices from disksreserved by the Disk Reservation agent.

system sysasystem sysbgroup Apache_1 (SystemList = { sysa ,sysb}AutoStartList = { sysa})

Apache myapacheWeb (

httpdDir = "/mnt/apache/bin"SecondLevelMonitor = 1ConfigFile = "/mnt/apache/conf/httpd.conf"HostName = "server1.example.com"Port = 80)

IP myapacheIP(Device = "eth0"Address="192.168.50.50"NetMask="255.255.255.0")

NIC myapacheNIC(Device="eth0"NetworkHosts={"172.29.9.178","172.29.9.179"})

Mount myapacheMnt(




MountPoint="/mnt/apache/"BlockDevice="/dev/sdd2"FSType = ext3

FsckOpt = "-y")

DiskReservation myapacheDiskRes(Disks = { "/dev/sdd" })

myapacheMnt requires myapacheDiskResmyapacheIP requires myapacheNICmyapacheWeb requires myapacheIPmyapacheWeb requires myapacheMnt

group Apache_2 (SystemList = { sysa,sysb}AutoStartList = { sysa})

Apache myapacheWeb2(httpdDir = "/mnt/apache1/bin"SecondLevelMonitor = 1ConfigFile = "/mnt/apache1/conf/httpd.conf"

HostName = "server2.example.com"Port = 8080)

IP myapacheIP2(Device = "eth1"Address="192.168.60.50"NetMask="255.255.255.0")

NIC myapacheNIC2(Device="eth1")

Mount myapacheMnt2(MountPoint="/mnt/apache1/"BlockDevice="/dev/sdc3"FSType = ext3FsckOpt = "-y"


172



)

DiskReservation myapacheDiskRes2(

Disks = { "/dev/sdc" })

myapacheMnt2 requires myapacheDiskRes2myapacheIP2 requires myapacheNIC2myapacheWeb2 requires myapacheIP2myapacheWeb2 requires myapacheMnt2

Sample main.cf fileA sample main.cf file follows:

include "types.cf"

cluster Cluster1 (UserNames = { admin = xxxxxx }

)

system SystemA (

)system SystemB ()

group Web1 (SystemList = { SystemA = 0, SystemB = 1 }

)

DiskGroup Web1_dg (DiskGroup = web1

)

Volume Web1_vol (DiskGroup = web1Volume = volweb1

)

IP Web1_ip (Device = eth0




Address = "10.212.88.220"NetMask = "255.255.254.0"

)

Mount Web1_mnt (MountPoint = "/apache/srvr01"BlockDevice = "/dev/vx/dsk/web1/volweb1"FSType = vxfsFsckOpt = "-y"

)

NIC Web1_nic (Device = eth0

)

Apache Web1_http (HostName = spartanPort = 80SecondLevelMonitor = 1SecondLevelTimeout = 25httpdDir = "/apache/srvr01/bin"EnvFile = "/apache/srvr01/bin/envvars"PidFile = /apache/srvr01/log/httpd.pid"ConfigFile = "/apache/srvr01/conf/httpd.conf"

IntentionalOffline = 1)

Web1_ip requires Web1_nicWeb1_mnt requires Web1_volWeb1_vol requires Web1_dgWeb1_http requires Web1_ipWeb1_http requires Web1_mnt

Basic IPv6 configuration

The following is a basic IPv6 configuration for the resource.

group ipv6group (SystemList = { sysA = 0, sysB = 1 })

Apache ipv6group_apache_res (HostName = "fd4b:454e:205a:110:211:25ff:fe7e:118"PidFile = "/myapache/apache/logs/httpd.pid"


174



httpdDir = "/myapache/apache/bin"ConfigFile = "/myapache/apache/conf/httpd.conf"ResLogLevel = TRACE

SecondLevelTimeout = 20IntentionalOffline = 1)

DiskGroup ipv6group_dg_res (DiskGroup = dg01)

IP ipv6group_ip_res (

Device = eth0

Address = "fd4b:454e:205a:110:211:25ff:fe7e:118"PrefixLen = 64)

Mount ipv6group_mnt_res (MountOpt = rwFsckOpt = "-n"BlockDevice = "/dev/vx/dsk/dg01/vol01"MountPoint = "/myapache/apache"FSType = vxfs

)

NIC ipv6group_nic_res (

Device = eth0

)Volume ipv6group_vol_res (

Volume = vol01DiskGroup = dg01

)

ipv6group_apache_res requires ipv6group_mnt_resipv6group_apache_res requires ipv6group_ip_resipv6group_mnt_res requires ipv6group_vol_resipv6group_vol_res requires ipv6group_dg_resipv6group_ip_res requires ipv6group_nic_res




Application agentTheApplication agent brings applications online, takes themoffline, andmonitors

their status. Use it to specify different executables for the online, offline, andmonitor routines for different programs. The executables can be on local storageor shared storage. You can use this agent to provide high availability forapplications that do not have bundled, enterprise, or custom agents.

An application runs in the default context of root. Specify the user name to runan application in a user context.

You can monitor the application in the following ways:

■ Use the monitor program

■ Specify a list of processes

■ Specify a list of process ID files

■ Any combination of the above

This agent is IMF-aware and uses asynchronous monitoring framework (AMF)kernel driver for IMF notification. For more information about the IntelligentMonitoring Framework (IMF) and intelligent resource monitoring, refer to theVeritas Cluster Server Administrator ’ s Guide.

PreventionOf Concurrency Violation (ProPCV) canbe enabledto prevent an onlineresource on a node from coming online on another node, outside of VCS control,in the same cluster. In that, ProPCV prevents the execution of StartProgram andprocesses that areconfiguredin MonitorProcesses on theoffline node. This actionprevents data corruption of resources and detects concurrency violation at anearly stage. The attribute can only be set for a local failover type group. To enablethis feature you need to set the ProPCV attribute value to 1. For more informationabout ProPCV, refer to the Veritas Cluster Server Administrator ’ s Guide.

High availability fire drill for Application agentThehigh availability firedrill detects discrepancies between theVCSconfigurationand the underlying infrastructure on a node. These discrepancies might prevent

a service group from going online on a specific node. For Application resources,the high availability fire drill checks for:

■ The availability of the specified program and execution permissions for thespecified program (program.vfd)

■ The existence of the specified user on the host (user.vfd)

■ The existence of the same binary on all nodes (cksum.vfd)

For more information, refer to the Veritas Cluster Server Administrator ’ s Guide .

Service and application agentsApplication agent

176



Dependencies for Application agentNo fixed dependency exists for Application agent.

Depending on how you plan to use it, an Application type of resource can dependon IP and Mount resources. Alternatively, instead of the IP resource you can alsouse the IPMultiNIC resource.

Figure 5-2 Sample service group that includes an Application resource

Application

IP Mount

NIC Volume

DiskGroup

Agent functions for Application agent

Runs the command or script that you specify in the value of theStartProgram attribute. Runs the command with the specified

parameters in the context of the specified user.To bring the resource online, the agent function performs thecommand:

su [-] user -c command_to_online_resource

Online

Runs the command or script that you specify in the value of theStopProgram attribute. Runs the command with the specifiedparameters in the context of the specified user.

To take the resource offline, the agent function performs thecommand:

su [-] user -c command_to_offline_resource

Offline




If youspecifytheMonitorProgram attribute, theagent executestheuserdefinedMonitorProgramin theuser-specifiedcontext.If you specify the PidFiles attribute, the routine verifies that

the process ID that is found in each listed file is running. If youspecify theMonitorProcessesattribute, theroutineverifies thateach listed process is running in the context of the user youspecify.

Useanycombinationamong these attributes (MonitorProgram,PidFiles, or MonitorProcesses) to monitor the application.

If any of the processes that are specified in either PidFiles orMonitorProcesses isdetermined notto be running, themonitorreturns OFFLINE. If the process terminates ungracefully, themonitor returns OFFLINE and failover occurs.

To monitor the resource, the agent function performs thecommand:

su [-] user -c command_to_monitor_resource

Monitor

Initializes the agent to interface with the asynchronousmonitoring framework (AMF)kernel driver. This function runswhen the agent starts up.

imf_init

Gets notification about resource state changes. This functionruns after theagent initializes with theAMFkernel driver. Theagent continuously waits for notification and takes action on

the resource upon notification.

imf_getnotification

Registers the resource entities, which the agent must monitor,with theAMF kernel driver. Forexample,the function registersthe PID for online monitoring of a process. This function runsfor each resource after the resource goes into steady state(online or offline). The Application agent uses IMF for theprocesses configured with PidFiles and the MonitorProcessesattribute.

imf_register


178



Terminatesprocesses specified in PidFiles or MonitorProcesses.Ensures that only those processes (that are specified in theMonitorProcesses attribute) running with the user ID specified

in the User attribute are killed. If the CleanProgram is defined,the agent executes the CleanProgram.

To forcefully stop the resource, the agent function performsthe command:

su [-] user -c command_to_clean_resource

Note that the agent uses the su - option only when theattribute UseSUDash is enabled (1). The UseSUDash attributeis disabled (0) by default.

Clean

The various functions of the action entry point are as follows:

■ program.vfdChecks the availability of the specified program and theexecution permissions for the specified program.

■ user.vfdChecks the existence of the specified user on the host.

■ cksum.vfdChecks the existence of the same binary on all nodes.

■ propcv[For internaluseonly]Invokes theAMF call with argumentsto decidewhether toallowor prevent processesfrom starting

for an application resource, outside the VCS control, in thecluster. The StartProgram and the processes configuredunder MonitorProcesses, registered with AMF for offlinemonitoring,are prevented fromstartingon theofflinenode.This helps prevent concurrency violation at an early stage.

■ getcksumReturns the checksum of the specified program

Action

State definitions for Application agent

Indicates that all processes that are specified in the PidFiles and theMonitorProcesses attribute are runningandthat theMonitorProgramreturns ONLINE.

ONLINE

Indicates that at least one process that is specified in the PidFilesattribute or MonitorProcesses is not running, or that theMonitorProgram returns OFFLINE.

OFFLINE




Indicates an indeterminable application state or invalidconfigurationor that the required attributes have not been configured.

UNKNOWN

Indicates that the process has terminated unexpectedly orMonitorProgram returns “ offline ” unexpectedly.

FAULTED

Attributes for Application agentTable 5-4 Required attributes for Linux


The executable which starts the application. The attribute specifies

the complete path of the executable. Applicable command linearguments follow the name of the executable. The executable pathand argument have spaces separating them. This executable can beon a local storage or on a shared storage.

For example, if the attribute for StartProgram is

/usr/sbin/vxnotify -g dg00 -m >> /var/log/vxnotify.log

(and vxnotify is blocking command) set it like:

/usr/sbin/vxnotify -g dg00 -m >> /var/log/vxnotify.log &

Note: Do not use the opening and closing ({ }) brace symbols in thisstring.


Example: "/usr/sbin/sample_app start"

StartProgram

The executable which stops the application. The Attribute specifiesthe complete path of the executable. Applicable command linearguments follow the name of the executable. The executable pathand argument have spaces separating them. This executable can beon a local storage or on a shared storage.

Note: Do not use the opening and closing ({ }) brace symbols in thisstring.


Example: "/usr/sbin/sample_app stop"

StopProgram


180



Table 5-4 Required attributes for Linux (continued)


See Table 5-5 on page 181.At least one of thefollowingattributes:

■ MonitorProcesses■ MonitorProgram■ PidFiles

Table 5-5 Optional attributes for Linux


Theexecutable which forciblystops the application. TheAttributespecifiesthecompletepath of theexecutable.Applicablecommandline arguments follow the name of the executable. The executablepath and argument have spaces separating them. This executablecan be on a local storage or on a shared storage.

Note: Symantec recommends to have the CleanProgram on thelocal storage so that in case of loss of storage connectivity VCScan take appropriate action to stop the application.


Example: "/usr/sbin/sample_app stop"

CleanProgram

A list of processes that you want monitored and cleaned. Eachprocess name is thename of an executable. Qualify theexecutablename with its complete path if the path starts the executable..

The process name must be the name that the ps -ef commanddisplays for the process.


Example: "nmbd"

MonitorProcesses






The executable which monitors the application. The Attributespecifiesthecompletepath of theexecutable.Applicable commandline arguments follow the name of the executable. The executablepath and argument have spaces separating them. This executablecan be on a local storage or on a shared storage.

MonitorProgramcan return the following VCSAgResState values:OFFLINE value is 100 or 1; ONLINE values range from 101 to 110or 0 (depending on the confidence level); 110 equals confidencelevel of 100%. Any other value = UNKNOWN.

Note: Do not use the opening and closing ({ }) brace symbols inthis string.

If MonitorProgram is configured and not available, then resourcestate will be:

■ OFFLINE if the resource was in OFFLINE state and not waitingfor any action.

■ UNKNOWN if the resource was in any other state or waitingfor some action.


Example: "/usr/sbin/sample_app_monitor all"

MonitorProgram

A list of PID (process ID) files that contain the PID ofthe processesthat you want monitored and cleaned. These are applicationgenerated files. Each PID file contains onemonitoredPID. Specifythe complete path of each PID file in the list.

The process ID can change when the process restarts. If theapplication takes time to update the PID file, the agent ’ s Monitorfunction mayreturnan incorrect result. If incorrect results occur,increase the ToleranceLimit in the resource definition.


PidFiles


182





The user ID for running StartProgram, StopProgram,MonitorProgram, and CleanProgram. The processes that arespecified in the MonitorProcesses list must run in the context of thespecifieduser. Monitor checksthe processes to make sure theyrun in this context.


Note: If configured user doesnotexistthen theresourcestatewillbe UNKNOWN.

Note: If home directory is not set for a configured user then theresource state will be UNKNOWN.

Default: root

Example: user1

User

The environment file that should get sourced before running anyof the StartProgram, StopProgram, MonitorProgram orCleanProgram.


Default: ""

Note: Please make sure that the EnvFile adheres the default shell

syntax of the configured user.Example: /home/username/envfile

EnvFile

When the value of this attribute is 0, the agent performs an suuser command before it executes the StartProgram, theStopProgram, the MonitorProgram, or the CleanProgram agentfunctions.

When the value of this attribute is 1, the agent performs an su -user command before it executes the StartProgram, theStopProgram, the MonitorProgram or the CleanProgram agentfunctions.


Default: 0

Example: 1

UseSUDash




Resource type definition for Application agent

type Application (static keylist SupportedActions = { "program.vfd","user.vfd", "cksum.vfd", getcksum, propcv }

static str ArgList[] = { User, StartProgram, StopProgram,CleanProgram, MonitorProgram, PidFiles, MonitorProcesses,EnvFile, UseSUDash, State, IState }

static int IMF{} = { Mode = 3, MonitorFreq = 1,RegisterRetryLimit = 3 }

static str IMFRegList[] = { MonitorProcesses, User, PidFiles,MonitorProgram, StartProgram }

str User = root

str StartProgramstr StopProgramstr CleanProgramstr MonitorProgramstr PidFiles[]str MonitorProcesses[]str EnvFileboolean UseSUDash = 0

)

Notes for Application agent

Using Application agent with IMFIntelligent monitoring is supported for the Application agent only under specificconfigurations. The complete list of such configurations is provided in thefollowing table:

Table 5-6

IMF MonitoringModePidFilesMonitorProcessesMonitorProgram

Not ApplicableNot ConfiguredNot ConfiguredNot Configured

Online, OfflineConfiguredNot ConfiguredNot Configured

Online, OfflineNot ConfiguredConfiguredNot Configured

Online, OfflineConfiguredConfiguredNot Configured


184



Table 5-6 (continued)

IMF MonitoringMode

PidFilesMonitorProcessesMonitorProgram

Offline OnlyNot ConfiguredNot ConfiguredConfigured

Offline OnlyConfiguredNot ConfiguredConfigured

Offline OnlyNot ConfiguredConfiguredConfigured

Offline OnlyConfiguredConfiguredConfigured

Note: When you do not configure MonitorProcesses, IMF monitors only theStartProgram on the offline node. Hence, the MonitorFreq of IMF attribute must

be set to 1 so that IMF monitors the resource on the offline node every monitorcycle.

Note: For a resource, if a PID file configured in the PidFiles attribute and a processconfiguredin theMonitorProcesses attributehave thesame process ID (PID), thenthe resource fails to register to IMF.

When multiple processes are configured under the MonitorProcesses attributeand only some of them are running, offline registration with IMF fails repeatedlyuntil RegisterRetryLimit is reached. In such a scenario, IMF cannot determinewhen the resource goes ONLINE and the agent monitors the resource in thetraditional way.

Using Application agent with ProPCVProPCV functionality prevents the StartProgramand binary-based processes thatare configured under MonitorProcesses from executing on the offline node. Thisaction detects concurrency violation at an early stage in the cycle. However,ProPCV does not prevent script-based processes that are configured underMonitorProcesses from executing on theoffline node. Considerations forProPCV

to function:■ Youmust runtheStartProgramwith thesame orderofargumentsasconfigured

in the StartProgram attribute. If you change the order of arguments, ProPCVdoes not prevent theexecutionof StartProgram. This causesdelayin detectingconcurrency violation.For example, a single command can be run in multiple ways:

/bin/tar -c -f a.tar




/bin/tar -f a.tar -c

So, ProPCV does not function if you run the command in a way that is notconfigured in the StartProgram attribute.

■ You must start the StartProgram by using the commands or the way specifiedin StartProgram attribute. But if you use another way or command to start theprogram that is not specified in the attribute, ProPCV does not prevent thestartup of the program. This causes delay in detecting concurrency violation.

■ If the StartProgram is a script, do not change the interpreter path in the scriptfile after the StartProgram is registered for offline monitoring. Else, ProPCVmay not function for the StartProgram.

■ You must not append the StartProgram attribute with the special character& . For example, '/app/start.sh &'.

Requirement for programsThe programs specified in StartProgram, StopProgram, MonitorProgram,CleanProgramshould not continuously write to STDOUTor STDERR. If required,please redirect STDOUT and STDERR to some file.

Requirement for default profileThe default profile of configured user should not have any blocking commandsuch as bash or any other command such as exec that changes the behavior of the shell. This may lead to unexpected behavior.

Sample configurations for Application agentThe sample configurations for this agent follow:

Configuration 1 for Application agentIn this example, you configure the executable sample_app as StartProgram andStopProgram, with start and stop specified as command line argumentsrespectively. Configure the agent to monitor two processes: a process that theapp.pid specifies and the process sample_app.

Application samba_app (User = "root"StartProgram = "/usr/sbin/sample_app start"StopProgram = "/usr/sbin/sample_app stop"PidFiles = { "/var/lock/sample_app/app.pid" }


186



MonitorProcesses = { "sample_app" })

Configuration 2 for Application agentIn this example, since no user is specified, it uses the root user. The executablesample_app starts andstops the application using start and stop as the commandline arguments. The executable sample_app_monitor monitors the applicationand uses all as its command line argument. The agent also monitors thesample_app1 and sample_app2 processes.

Application samba_app2 (StartProgram = "/usr/sbin/sample_app start"StopProgram = "/usr/sbin/sample_app stop"

CleanProgram = "/usr/sbin/sample_app force stop"MonitorProgram = "/usr/local/bin/sample_app_monitor all"MonitorProcesses = { "sample_app1", "sample_app2" })

Debug log levels for Application agentThe Application agent uses the following debug log levels:


CoordPoint agentUse the Coordination Point (CoordPoint) agent to monitor the registrations onthe different coordination points on each node.

In addition, theCoordPointagent monitors changes to theCoordinator Disk Groupconstitution, such as when a disk is accidently deleted from the Coordinator DiskGroup or if the VxVM private region of a disk is corrupted.

The agent performs detailed monitoring on the CoordPoint resource. You cantune the frequency of the detailed monitoring with the LevelTwoMonitorFreqattribute. For example, if you set this attribute to 5, the agent monitors theCoordinator Disk Group constitution in every fifth monitor cycle.

The CoordPoint agent is a monitor-only agent that runs on each node within theclient cluster. It can monitor Coordination Point (CP) servers and SCSI-3 disks.

Service and application agentsCoordPoint agent



Coordination Point server as a coordination pointWhen you have configured a CP server as a coordination point, the CoordPointagent performs the following tasks:

■ Confirms that the CP server coordination point can communicate with theclient cluster.

■ Validates the node registrations in the CP server database using the cpsadmcommand.

SCSI-3 based disk as a coordination pointIn case the coordination point is a SCSI-3 based disk, the CoordPoint agent usesthe vxfenadm command to confirm that the registered keys on the disk are intact.

The Monitor agent function contains the monitoring functionality for SCSI-3disks and CP servers.

If the agent detects an anomaly, the agent reports it to you so you can repair thecoordination point. You may have to perform an online coordinator pointreplacement procedure if the problem is isolated to the keys registered.

Note: The CoordPoint agent that runs on a given client cluster node monitors thekeys for coordination points visible to that node alone.

For important information about this agent, refer to:

See “ Notes for the CoordPoint agent ” on page 191.

DependenciesNo dependencies exist for the CoordPoint resource.


188



Agent functions

Enables theCoordPoint agent tovalidate thenode registrations in the

coordination points and confirms that the coordination points areaccessible. In addition, enables the agent to monitor disks in theCoordinator Disk Group. Specifically, if a disk is deleted from thediskgroup or the VxVM private region of a disk is corrupted.

CoordPoint resources are persistent, which means that they cannotbe brought online or taken offline. They can only monitor thecoordination point registrations. For this reason, the service groupthat contains the CoordPoint resource appears to be offline after acommand such as hastatus -sum .

The CoordPoint agent also performs I/O fencing reporting activities.

See “

CoordPoint agent I/O fencing reporting activities”

on page 191.

Monitor

State definitions

Indicates that the CoordPoint resource is working.ONLINE

Indicates the agent cannot determine the coordination pointsresource's state. This state may be due to an incorrect configuration.

UNKNOWN

Indicates that CoordPoint resource is reported for one or more of thefollowing conditions:

■ The number of coordination points with missing keys (orregistrations) has exceeded the value of the FaultToleranceattribute.

■ The number of unreachable coordination points.■ Coordinator disks are deleted from the Coordinator Disk Group.■ Publiccharacterpathof a disk andthe devicepath that corresponds

to thedevice numberof that disk in the kernel driver donotmatch.

FAULTED






The FaultTolerance attribute determines when the CoordPoint agentdeclares that theregistrations on thecoordination pointsaremissingor connectivitybetween thenodesand thecoordinationpoints is lost.

If the number of coordination points with missing keys (orregistrations) and or the number of unreachable coordination pointsexceeds the value of the FaultTolerance attribute, then the agentreports FAULTED.

Set the value of this attribute depending on your own configurationrequirements. For example, if the FaultTolerance value is set to 1,then the CoordPoint agent reports FAULTED if it sees 2 or morenumber of coordinator points withmissing keys(or registrations) andor the number of unreachable coordination points.

Change the value of the FaultTolerance attribute either before theCoordPoint agent starts to monitor or while the CoordPoint agent ismonitoring. If the attribute is set while the CoordPoint agent ismonitoring, then the CoordPoint agent reads the new value in thenext monitor cycle.

To view the current FaultTolerance value, enter the following

command:

# hares -display coordpoint -attribute FaultTolerance


Default: "0"

FaultTolerance

Resource type definitiontype CoordPoint (

static str ArgList[] = { FaultTolerance }

static int InfoInterval = 300static int OfflineMonitorInterval = 60static str Operations = Noneint FaultTolerance

)


190



Notes for the CoordPoint agentThe notes are as follows:

CoordPoint agent I/O fencing reporting activitiesTheCoordPoint agent alsoperforms the following I/Ofencing reporting activities:

■ Checks to determine if I/O fencing is running.If I/O fencing is not running, then the CoordPoint agent reports failure.

■ Checks the mode of fencing operation. I/O fencing can operate in one of thefollowing three modes:

■ SCSI-3 mode: If I/O fencing runs in SCSI-3 mode, then theCoordPointagentcontinues to monitor.

■ Customized mode: If I/O fencing runs in Customized Fencing mode, thenthe CoordPoint agent continues to monitor.

■ Disabled mode: If I/O fencing runs in disabled mode, no action is required.The CoordPoint agent returns success.

AutoStartList attributeAutoStartList is a service groupattribute that needs to be populated with a systemlist. The VCS engine brings up the specified service group on the nodes in the list.

AutoStartList is not a required attribute for the service group that contains theCoordPoint resource. The CoordPoint resource is a persistent resource and whena service group is configured with this type of resource, it cannot be broughtonline.

Specifying the AutoStartList with a system list does not change the behavior of the service group. The service group will be reflected in OFFLINE status itself,irrespective of the AutoStartList attribute.

Detailed monitoring for the Coordpoint resourceThe agent fetches disk names and unique identifiers from the kernel driver forI/O fencing. It runs a series of commands on the disks for information such asdisk access name. It checks for disks that are no longer part of the CoordinatorDisk Group. It also compares thepubliccharacter path of thedisks with thedevicepath stored in the kernel driver. The agent faults the resource when any of thechecks fail.




Sample configurationIn this example, the coordination point agent type resource is configured withthe value of the FaultTolerance attribute set to 0. At this value setting, theCoordPoint agent reports FAULTED, when the agent determines that at least onecoordination point has keys (or registrations) missing and or one coordinationpoint is not reachable.

The following is an example service group (vxfen) extracted from a main.cf file:

group vxfen (SystemList = { sysA = 0, sysB = 1 }AutoFailOver = 0Parallel = 1AutoStartList = { sysA, sysB }

)CoordPoint coordpoint (

FaultTolerance=0LevelTwoMonitorFreq = 5)

// resource dependency tree//// group vxfen// {// CoordPoint coordpoint// }

Debug log levelsThe CoordPoint agent uses the following debug log levels:

DBG_10

KVMGuest agentThe KVMGuest agent monitors a Virtual Machine (KVM guest) created in KVM

environment or Red Hat Enterprise Virtualization Environment.Agent brings virtual machines online and takes it offline. The KVMGuest agentuses virsh commands to manage the virtual machine in KVM environment andRepresentational state transfer (REST) APIs to manage the virtual machines inRed Hat Enterprise Virtualization environment.

You can use this agent to make a virtual machine highly available and to monitorit.

Service and application agentsKVMGuest agent

192



Veritas Cluster Server supports guest virtual machines created on Red HatEnterprise Linux 6 and SuSEEnterprise Linux 11SP2 and virtualmachines createdin Red Hat Enterprise Virtualization 3.0.

For important information on this agent, refer to:See “ Notes for KVMGuest agent ” on page 197.

Dependencies for KVMGuest agentKVM Environment: The KVMGuest resource may depend on the Mount resourceof CFSMount resource. The dependency is based on where the virtual machineimage is located.

RHEV Environment: The KVMGuest resource may not depend on any resource.

Figure 5-3 Sample service group that includes a KVMGuest resource

Agent functions for KVMGuest agent

KVM environment: Agent uses the virsh start command to starttheguest virtual machine. When theresource is configured to definethe guest configuration, agent uses the virsh define command todefine the virtual machine while bringing it online.

RHEV environment: Agent uses the REST APIs to start the virtualmachine.

Online

KVM environment: Agent uses the virsh shutdown command toshutdown the guest virtual machine. If the SyncDir attribute isconfigured to synchronize the guest virtual machine configurationfile, then the configuration file is copied to the location configuredas a SyncDir attribute.

RHEV environment:Agentuses theRESTAPIsto shutdownthevirtualmachine.

Offline




KVM environment: Agent uses the virsh domstate command todetermine the status of the guest virtual machine.

RHEV environment: Agent uses the REST APIs to get the status of the virtual machine.

Monitor

KVM environment: Agent uses the virsh destroy command toforcefullyshutdowntheguest virtual machine. If theSyncDir attributeis configured to synchronize theguest virtual machine configurationfile, then the configuration file is copied to the location configuredas a SyncDir attribute.

RHEV environment: Agent uses REST APIs to stop the virtualmachine.

Clean

Starts the vcskvmguestd process which is used to detect the virtual

machine migration.

Open

Stops the vcskvmguestd processClose

Stops the vcskvmguestd processShutdown

checks the value of DelayAfterGuestOnline andDelayAfterGuestOffline attribute. If it is less than the default valuethen the warning message is logged.

attr_changed

guestmigrated

vcskvmguestd process detects the virtual machine migration and

executes this actionentry point to create themigrationstate file andinitiate the resource monitor using the hares -probe command.

Action

State definitions for KVMGuest agent

Indicates that the virtual machine is running.ONLINE

Indicates that the virtual machine has stopped.OFFLINE

Indicates that the virtual machine has failed to start orunexpectedly stopped.

FAULTED

Indicates that the problem exists with the configuration or theability to monitor the resource.

UNKNOWN

Indicates that the virtual machine is either migrated to anotherphysical host or the guest virtual machine is intentionallysuspended by the administrator.

INTENTIONALOFFLINE


194



Attributes for KVMGuest agentTable 5-8 Required attribute


The name of the virtual machine created using the KVMhypervisor or RHEV-M.


Example: vm1

GuestName

Table 5-9 Optional attribute


Defines the maximum time in seconds that the virtual machinetakes to start. You can modify this attribute as required.


Default: 5

Example: 10

DelayAfterGuestOnline

Defines the maximum time in seconds that the virtual machinetakes to shut down. You can modify this attribute as required.


Default: 30

Example: 40

DelayAfterGuestOffline

Specifies the absolute path of a directory used to synchronizethedefault configurationdirectory /etc/libvirt/qemu/on all thecluster nodes. If this attribute is configured, the online entrypoint uses the guest configuration file in the specified path todefine the guest. This directory must be on the shared storage.

This attribute is valid in KVM environment only.


Example: /kvmguest_syncdir

SyncDir




Table 5-9 Optional attribute (continued)


Specifies the absolute path of the guest configuration file. Youcan use this attribute as an alternative to SyncDir. If youconfigurethisattribute,ensure that the guest configuration fileis available on all the cluster nodes and on same path. VCS usesthis path to define the guest on each node. If both SyncDir andGuestConfigFilePath are configured, then SyncDir is preferredover GuestConfigFilePath.

This attribute is valid in KVM environment only.


Example: /kvmguest/kvmguest1.xml

GuestConfigFilePath

Specifies information about the RHEV environment

The Keys associated with this attribute are:

■ Enabled.

Specifies the virtualization environment. If value is:■ 0 : KVM Environment■ 1 : RHEV Environment■ Default : 0

■ URL: Specifies the RHEV-M URL that can be used for RESTAPI communication.

Example: https://rhevm-server.symantec.com:8443■ User: Specifies the RHEV-M User that can be used for REST

API communication.■ Password:Specifies theencryptedpasswordof RHEV-M User.

The password should be encrypted using vcsencryptcommand.

■ Cluster: Specifies the name of the RHEV-M cluster, to whichthe VCS host belongs.

RHEVMInfo


DescriptionInternal attribute

For information on how to use the IntentionalOffline resourcetype attribute, refer to the Veritas Cluster ServerAdministrator'sGuide.

IntentionalOffline

This is an internal attribute.CEInfo


196



Table 5-10 Internal attributes (continued)

DescriptionInternal attribute

The ResyncVMCfg attribute is setby the havmconfigsync utility.If this attribute is set, the agent redefines the virtual machineconfiguration if it already exists using the SyncDir attribute. If the SyncDir attribute is not set, GuestConfigFilePath attribute isused.

Note: You must not set the ResyncVMCfg attribute manually.

ResyncVMCfg

Resource type definition for KVMGuest agent

type KVMGuest (static int IntentionalOffline = 1static keylist SupportedActions = { "guestmigrated","vmconfigsync" }static keylist RegList = { "GuestName", "DelayAfterGuestOnline","DelayAfterGuestOffline", "RHEVMInfo" }static str ArgList[] = { GuestName, DelayAfterGuestOnline,DelayAfterGuestOffline, SyncDir, GuestConfigFilePath, CEInfo,RHEVMInfo, ResyncVMCfg }str CEInfo{} = { Enabled=0, CESystem=NONE, FaultOnHBLoss=1 }

str RHEVMInfo{} = { Enabled=0, URL=NONE, User=NONE,Password=NONE, Cluster=NONE }str GuestNameint DelayAfterGuestOnline = 5int DelayAfterGuestOffline = 30str SyncDirstr GuestConfigFilePathboolean ResyncVMCfg = 0

)

Notes for KVMGuest agentThe KVMGuest agent has the following notes:

■ Support forguestscreated on RedHatEnterprise Linux 6 andSuSE EnterpriseLinux 11 SP2

■ Storage and network configurations

■ Guest live migration




■ Managing virtual machines in RHEV environment

Support for guests created on Red Hat Enterprise Linux 6 andSuSE Enterprise Linux 11 SP2KVMGuest agent supports only those guests that can be created on Red HatEnterprise Linux 6 and SuSE Enterprise Linux 11 SP2.

Storage and network configurationsIf you have specified customized network and storage configurations for anyguest, make sure that same configurations are available on all the cluster nodes.

Guest live migrationFor migrating the guest from one node to another node, make sure that the guestimage is available on both the nodes simultaneously at the same location. Also,KVMrequires that theguest networkandstorage configuration shouldbe identicalon source and destination node. For details, refer to the Redhat Enterprise Linux Virtualization Guide .

Managing virtual machines in RHEV environmentRefer to the Veritas Storage Foundation and High Availability SolutionsVirtualization Guide.

Managing ISO image in SuSE KVMBy default, KVM guest virtual machines created on SuSE requires the ISO imageto be alwaysavailable while starting thevirtualmachine. After thevirtualmachineis created, you can modify the virtual machine configuration to remove thereference to the ISO image. Alternatively, you can configure the VCS Mountresource to make the ISO image available on a cluster node.

Sample configurations for KVMGuest agentFollowing are few sample configurations for managing virtual machines in KVMenvironment.

Sample Configuration 1 for KVMGuest agentIn this example, the native LVM volumes are used to store the guest image.

group kvmtest1 (SystemList = { north = 0, south = 1 }


198



)Volume vol1 (Volume = kvm_vol

DiskGroup = kvm_dg)DiskGroup dg1 (DiskGroup = kvm_dg)res1 requires mnt1mnt1 requires vol1vol1 requires dg1

Sample Configuration 3 for KVMGuest agent

In this example, the native CVM-CFS is used to store the guest image.

group cvm (SystemList = { kvmpm1 = 0, kvmpm2 = 1 }AutoFailOver = 0Parallel = 1AutoStartList = { kvmpm1, kvmpm2 })

CFSMount cfsmount (MountPoint = "/cfsmount"

BlockDevice = "/dev/vx/dsk/cfsdg/cfsvol")

CFSfsckd vxfsckd ()

CVMCluster cvm_clus (CVMClustName = kvmcfsCVMNodeId = { kvmpm1 = 0, kvmpm2 = 1 }CVMTransport = gab

CVMTimeout = 200)

CVMVolDg cfsdg (CVMDiskGroup = cfsdgCVMVolume = { cfsvol }CVMActivation = sw)


200



CVMVxconfigd cvm_vxconfigd (Critical = 0CVMVxconfigdArgs = { syslog }

)

cfsmount requires cfsdgcfsmount requires cvm_cluscvm_clus requires cvm_vxconfigdvxfsckd requires cvm_clus

Sample configurations for RHEV environmentFollowing arefew sampleconfigurations formanaging virtual machines in RHEVenvironment.

Sample Configuration 1 for KVMGuest agent

group rhev_grp1 (SystemList = { north = 0, south = 1 })KVMGuest kvmres1 (RHEVMInfo = { Enabled = 1,URL = "https://rhevm-server.symantec.com:8443",

User = admin,Password = bncNfnOnkNphChdHe,Cluster = dc2_cluster1 }GuestName = rhevvm1DelayAfterGuestOnline = 20DelayAfterGuestOffline = 35)

Sample Configuration 2 for multiple KVMGuest resources

group rhev_grp1 (SystemList = { north = 0, south = 1 })KVMGuest kvmres1 (RHEVMInfo = { Enabled = 1,URL = "https://rhevm-server.symantec.com:8443",User = admin,Password = bncNfnOnkNphChdHe,




Cluster = dc2_cluster1 }GuestName = rhevvm1DelayAfterGuestOnline = 20

DelayAfterGuestOffline = 35)

group rhev_grp2 (SystemList = { north = 0, south = 1 })KVMGuest kvmres2 (RHEVMInfo = { Enabled = 1,URL = "https://rhevm-server.symantec.com:8443",User = admin,Password = bncNfnOnkNphChdHe,

Cluster = dc2_cluster1 }GuestName = rhevvm2DelayAfterGuestOnline = 20DelayAfterGuestOffline = 35)

Sample Configuration for SuSE KVM

group kvmgrp (

SystemList = { north = 0, south = 1 })KVMGuest kvmres1 (

GuestName = kvmguest1DelayAfterGuestOnline = 10DelayAfterGuestOffline = 30)

requires group mntgrp online local firm

group mntgrp (SystemList = { north = 0, south = 1 }AutoFailOver = 0Parallel = 1AutoStartList = { north, south })

Mount mntres1 (MountPoint = "/os_iso_image"


202



BlockDevice = "nfsserver:/os/suse"FSType = nfsFsckOpt = "-y"

)

Debug log levels for KVMGuest agentThe DiskGroup agent uses the following debug log levels:


Process agentThe Process agent starts, stops, and monitors a process that you specify. You canuse the agent to make a process highly available.

This agent is Intelligent Monitoring Framework (IMF)-aware and usesasynchronous monitoring framework (AMF) kernel driver for IMF notification.Formore information about IMFand intelligent resource monitoring, refer to theVeritas Cluster Server Administrator ’ s Guide.

The agent does an exact match of configured PathName and Arguments with theprocesses in the process table. To clarify, the processes that ps command listsshould have exact match of PathName and Arguments for all the configuredprocesses in the VCS configuration file /etc/VRTSvcs/conf/config/main.cf .

Note that the AMF kernel driver does not monitor kernel processes. Even if youhave enabled intelligent monitoringforProcess agent,youmust use the traditionalpoll-based monitoring to monitor kernel processes.

High availability fire drill for Process agentThehigh availability firedrill detects discrepancies between theVCSconfigurationand the underlying infrastructure on a node; discrepancies that might prevent aservice group from going online on a specific node.

For Process resources, the high availability fire drill checks for:

■ The existence of a binary executable for the specified process (program.vfd)

■ The existence of the same binary on all nodes (program.vfd)

For more information refer to the Veritas Cluster Server Administrator ’ s Guide .

Service and application agentsProcess agent



Dependencies for Process agentDepending on the context, this type of resource can depend on IP, IPMultiNIC,and Mount resources.

Figure 5-4 Sample service group for a Process resource

Process

IP Mount

NIC Volume

DiskGroup

Agent functions for Process agent

Starts a processin thebackgroundwith optionalarguments and priority in the specified usercontext.

Online

Terminates the process with a SIGTERM. If the

process does not terminate, a SIGKILL is sent.

Offline

Checks to see if the process is running byscanning the process table for the name of theexecutable pathname and argument list.

Monitor

Terminates all ongoing resource actions andtakes the resource offline, forcibly whennecessary.

Clean

Initializes the agent to interface with theasynchronous monitoring framework (AMF)

kerneldriver. This function runs when theagentstarts up.

imf_init

Gets notification about resource state changes.This function runs after the agent initializeswith the AMF kernel driver. The agentcontinuously waits for notification and takesaction on the resource upon notification.

imf_getnotification


204



Registersthe resource entities,whichthe agentmust monitor, with the AMF kernel driver. Forexample, the function registers the PID for

online monitoring of a process. This functionruns for each resource after the resource goesinto steady state (online or offline).

imf_register

State definitions for Process agent

Indicates that the specified process is running.

The agent only reports the process as online if the value configuredfor PathName attribute exactly matches the process listing from theps output along with the arguments.

ONLINE

Indicates that the specified process is not running.OFFLINE

Indicates that the process has terminated unexpectedly.FAULTED

Indicates that the agent can not determine the state of the process.UNKNOWN

Attributes for Process agentTable 5-11 Required attribute for Linux


Absolute path to access an executable program. This path includesthe program name. If a script controls the process, the PathNamedefines the complete path to the shell.

This attribute must not exceed 256 characters.


Example: "/usr/sbin/proc1"

PathName






Passes arguments to the process. If a script controls the process, thescript is passed as an argument. Separate multiple arguments with asingle space. A string cannot accommodate more than one spacebetween arguments, nor allow for leading or trailing whitespacecharacters.


Arguments

The file that contains the process ID for the monitoring process.Specify thePidFile attributefor themonitoring process to usethe Pid.Otherwise, to complete the monitoring process the agent uses the ps

output.Note that when youuse scripts, or other indirect mechanisms, to startprocesses, youmust set thePidFile attributeif theps output isdifferentfrom theconfigured valuesforthePathNameor Argumentsattributes.


Example: "/var/lock/sendmail.pid"

PidFile

Priority that the process runs. Priority values range between -20(highest) to +19 (lowest).


Default: 10

Priority

This attribute is the owner of the process. The process runs with thisuser ID.


Default: root

UserName

Resource type definition for Process agenttype Process (

static keylist SupportedActions = { "program.vfd", getcksum }static str ArgList[] = { PathName, Arguments, UserName,Priority, PidFile }static int IMF{} = { Mode = 3, MonitorFreq = 5,RegisterRetryLimit = 3 }str PathNamestr Argumentsstr UserName = root


206



DependenciesNo child dependencies exist for this resource.

Agent functions

Starts the process with optional arguments.Online

Checks to see if the process is alive by scanning the process table forthe name of the executable pathname and argument list.

Monitor


Clean

State definitionsIndicates that the specified process is running.

Theagent only reports the process as ONLINE if thevalue configuredfor PathName attribute exactly matches the process listing from theps output along with the arguments.

ONLINE

Indicates that the process has unexpectedly terminated.FAULTED

Indicates that the agent can not determine the state of the process.UNKNOWN

AttributesTable 5-13 Required attributes for Linux


Defines absolute path to access an executable program. This pathincludes the program name. If a process is controlled by a script, thePathName defines the complete path to the shell. The PathNameattribute must not exceed 256 characters.

The value configured for this attribute needs to match the processlisting from the ps output for the agent to display as ONLINE.


PathName

Service and application agentsProcessOnOnly agent

208





Passes arguments to the process. If a process is controlled by a script,the script is passed as an argument. Multiple arguments must beseparatedby a single space. A string cannot accommodate more thanone space between arguments, nor allow for leading or trailingwhitespace characters.


Example: "-bd -q30m"

Arguments

A flag that indicates whether monitor ignores the argument list.

■

If thevalueis 0, it checksthe process pathname andargument list.■ If the value is 1, it only checks for the executable pathname and

ignores the rest of the argument list.


Default: 0

IgnoreArgs

The file that contains the process ID for the monitoring process.Specify thePidFile attributefor themonitoringprocess to usethe Pid.Otherwise, to complete the monitoring process the agent uses the psoutput.

Note that when youuse scripts, or other indirect mechanisms, to startprocesses, you must set the PidFile attribute when the ps output isdifferent from theconfigured valuesfor thePathName or Argumentsattributes.


Example: "/var/lock/sendmail.pid"

PidFile

Prioritywith which theprocesswillrun. Priorityvaluesrange between-20 (highest) to +19 (lowest).


Default: 10

Priority

Owner of the process. The process runs with the user ID.


Default: root

UserName




Resource type definitiontype ProcessOnOnly (

static str ArgList[] = { PathName, Arguments, UserName,Priority, PidFile, IgnoreArgs }static str Operations = OnOnlystr PathNamestr Argumentsstr UserName = rootstr Priority = 10str PidFileboolean IgnoreArgs = 0

)

ProcessOnOnly agent usage notesThe ProcessOnOnly agent has the following notes:

■ Requirement for programs

Requirement for programsThe programs specified in PathName should not continuously write to STDOUTor STDERR. If required, please redirect STDOUT and STDERR to some other file.

Sample configurationsgroup VxSS (SystemList = { north = 0, south = 1 }Parallel = 1AutoStartList = { north, south }OnlineRetryLimit = 3OnlineRetryInterval = 120)Phantom phantom_vxss (

)ProcessOnOnly vxatd (IgnoreArgs = 1PathName = "/opt/VRTSat/bin/vxatd")

Debug log levelsThe ProcessOnOnly agent uses the following debug log levels:


210



DBG_1, DBG_4





212



Other applications with the name notifier can interfere with the NotifierMngragent. If notifier is startedoutside VCS control, VCS can only monitor the notifierprocess if its started with the absolute path. For example, use:

# /opt/VRTSvcs/bin/notifier -s m=xyz &

DependencyThe NotifierMngr resource can depend on the NIC resource.

Agent functions

Starts the notifier process with its required arguments.Online

VCSsends a SIGABORT. If theprocessdoesnot exit within one second,VCS sends a SIGKILL .

Offline

Monitors the notifier process.Monitor

Sends SIGKILL .Clean

State definitions

Indicates that the Notifier process is running.ONLINE

Indicates that the Notifier process is not running.OFFLINE

Indicates that the user did not specify the required attribute for theresource.

UNKNOWN

Infrastructure and support agentsNotifierMngr agent

214



AttributesTable 6-1 Required attributes for Linux


Specifies the machine names of the SNMP managers and the severitylevel of the messages to be delivered. The severity levels of messagesare Information , Warning , Error , and SevereError . Specifyinga given severity level for messages generatesdelivery of all messagesof equal or higher severity.

SnmpConsoles is a required attribute if SmtpServer is not specified;otherwise, SnmpConsoles is an optional attribute. Specify bothSnmpConsoles and SmtpServer if desired.


Example:

"172.29.10.89" = Error, "172.29.10.56" = Information

SnmpConsoles

Specifies the machine name of the SMTP server.

SmtpServer is a required attribute if SnmpConsoles is not specified;otherwise, SmtpServer is an optional attribute. You can specify bothSmtpServer and SnmpConsoles if desired.


Example: "smtp.example.com"

SmtpServer



Change this attribute if the VCS engine is listening on a portother than its default port.


Default: 14141

EngineListeningPort

Size of the VCS engine ’ s message queue. Minimum value is30.


Default: 30

MessagesQueue






Any valid, unused TCP/IP port number.Type and dimension: integer-scalar

Default: 14144

NotifierListeningPort

If this attribute is populated, all the notifications sent fromthe notifier (SMTPandSNMP) will be sent from the interfacehaving this IP address.

Note: Make sure that the SourceIP given in this attribute ispresent in the /etc/hosts file or is DNS-resolvable.


Example: "10.209.77.111"

NotifierSourceIP

Set to a valid email address, if you want the notifier to use acustom email address in the FROM: field.


Example: "[email protected]"

SmtpFromPath

Specifies the email address where SMTP sends informationandtheseverity level of the messages.The severity levelsof messages are Information,Warning, Error, andSevereError.

Specifying a given severity level for messages indicates thatall messages of equal or higher severity are received.

Note: SmtpRecipients is a required attribute if you specifySmtpServer.


Example:

"[email protected]" = SevereError,"[email protected]" = Warning

SmtpRecipients

Set to a valid email address, if you want the notifier to use acustom email address in the Return-Path: <> field.

If the mail server specified in SmtpServer does not supportSMTP VRFY command, then youneed to set theSmtpVrfyOff to 1 in order for the SmtpReturnPath value to take effect.


Example: "[email protected]"

SmtpReturnPath


216





This attribute represents the time in seconds notifier waitsfora response from themail serverfor theSMTP commandsit has sent to the mail server. This value can be increased if you notice that the mail server is taking a longer durationto reply back to the SMTP commands sent by notifier.


Default: 10

SmtpServerTimeout

Set thisvalue to 1 if your mail server doesnot support SMTPVRFY command. If you set this value to 1 , the notifier doesnot send a SMTP VRFY request to the mail server specifiedin SmtpServer attribute while sending emails.


Default: 0

SmtpServerVrfyOff

Specifies the community ID for the SNMP manager.


Default: public

SnmpCommunity

Port on the SNMP console machine where SNMP traps aresent.

If you specify more than one SNMP console, all consoles usethis value.


Default: 162

SnmpdTrapPort

Resource type definitiontype NotifierMngr (static int RestartLimit = 3

static str ArgList[] = { EngineListeningPort, MessagesQueue,NotifierListeningPort, NotifierSourceIP, SnmpdTrapPort,SnmpCommunity, SnmpConsoles, SmtpServer, SmtpServerVrfyOff,SmtpServerTimeout, SmtpReturnPath, SmtpFromPath, SmtpRecipients}int EngineListeningPort = 14141int MessagesQueue = 30int NotifierListeningPort = 14144




str NotifierSourceIPint SnmpdTrapPort = 162str SnmpCommunity = public

str SnmpConsoles{}str SmtpServerboolean SmtpServerVrfyOff = 0int SmtpServerTimeout = 10str SmtpReturnPathstr SmtpFromPathstr SmtpRecipients{})

Sample configurationIn the following configuration, the NotifierMngr agent is configured to run withtwo resource groups: NicGrp and Grp1. NicGrp contains the NIC resource and aPhantom resource that enables VCS to determine the online and offline status of the group. See the Phantom agent for more information on verifying the statusof groups that only contain OnOnly or Persistent resources such as the NICresource. You must enable NicGrp to run as a parallel group on both systems.

Grp1 contains the NotifierMngr resource (ntfr) and a Proxy resource (nicproxy),configured for the NIC resource in the first group.

In this example, NotifierMngr has a dependency on the Proxy resource.

Note: Only one instance of the notifier process can run in a cluster. The processcannot run in a parallel group.

TheNotifierMngrresource sets up notification forall events to theSNMP consolesnmpserv . In this example, only messages of SevereError level are sent to theSMTP server ( smtp.example.com ), and the recipient ( [email protected] ).

Configuration

Configuration for Linux follows:system north

system south

group NicGrp (SystemList = { north, south}AutoStartList = { north }


218



Parallel = 1)

Phantom my_phantom ()

NIC NicGrp_eth0 (Device = eth0

)group Grp1 (

SystemList = { north, south }AutoStartList = { north })

Proxy nicproxy(TargetResName = "NicGrp_eth0")

NotifierMngr ntfr (SnmpConsoles = { snmpserv = Information }SmtpServer = "smtp.example.com"SmtpRecipients = { "[email protected]" =SevereError }

)

ntfr requires nicproxy

// resource dependency tree//// group Grp1// {// NotifierMngr ntfr// {// Proxy nicproxy// }// }

Debug log levelsThe NotifierMngr agent uses the following debug log levels:

DBG_1




Proxy agentTheProxy agentmirrors thestate of another resource on a local or remotesystem.

It provides a means to specify andmodifyoneresourceandhave its state reflectedby its proxies. You can use the agent when you need to replicate the status of aresource.

A Proxy resource can only point to None or OnOnly type of resources, and canreside either in a failover or a parallel group. A target resource and its proxycannot be in the same group.

DependenciesNo dependencies exist for the Proxy resource.

Figure 6-1 Sample service group that includes a Proxy resource

App1

IP Mount

NIC Volume

DiskGroup

App2

IP Mount

Proxy Volume

DiskGroup

Agent functions

Determines status based on the target resource status.Monitor

Infrastructure and support agentsProxy agent

220



AttributesTable 6-3 Required attribute


Name of the target resource that the Proxy resource mirrors.

The target resource must be in a different resource group than theProxy resource.


Example: "nic1"

TargetResName

Table 6-4 Optional attribute


Mirrors the status of the TargetResName attribute on systems thattheTargetSysName variablespecifies. If thisattribute is notspecified,the Proxy resource assumes the system is local.


Example: "sysa"

TargetSysName

Resource type definitiontype Proxy (

static int OfflineMonitorInterval = 60static str ArgList[] = { TargetResName, TargetSysName,"TargetResName:Probed", "TargetResName:State" }static str Operations = Nonestr TargetResNamestr TargetSysName

)


Configuration 1

Proxy proxy1 (TargetResName = "nic1"

)




Configuration 2The proxy resource mirrors the state of the resource nic2 on sysa.

Proxy proxy1(TargetResName = "nic2"TargetSysName = "sysa"

)

Configuration 3The proxy resource mirrors the state of the resource mnic on the local system;note that target resource is in grp1, and the proxy is in grp2; a target resourceand its proxy cannot be in the same group.

group grp1 (SystemList = { sysa, sysb }AutoStartList = { sysa })

MultiNICA mnic (Device@sysa = { eth0 = "192.123.8.41", eth3 ="192.123.8.42" }Device@sysb = { eth0 = "192.123.8.43", eth3 ="192.123.8.43" }NetMask = "255.255.248.0"

NetworkHosts = { "192.123.10.129", "192.123.10.130" })

IPMultiNIC ip1 (Address = "192.123.10.177"MultiNICAResName = mnicNetMask = "255.255.248.0")

ip1 requires mnic

group grp2 (SystemList = { sysa, sysb }AutoStartList = { sysa })

IPMultiNIC ip2 (


222



Agent functions

Determines status based on the status of the service group.Monitor

Resource type definitiontype Phantom ()


Configuration 1

Phantom boo ()

Configuration 2Thefollowing example shows a complete main.cf, in which the FileNone resourceand the Phantom resource are in the same group.

include "types.cf"cluster PhantomClustersystem sysa ()system sysb ()group phantomgroup (SystemList = { sysa = 0, sysb = 1 }AutoStartList = { sysa }Parallel = 1)FileNone my_file_none (PathName = "/tmp/file_none"

)Phantom my_phantom ()// resource dependency tree//// group maingroup// {// Phantom my_Phantom

Infrastructure and support agentsPhantom agent

224



// FileNone my_file_none// }

RemoteGroup agentThe RemoteGroup agent establishes dependencies between applications that areconfigured on different VCS clusters. For example, you configure an Apacheresource in a local cluster, and a MySQL resource in a remote cluster. In thisexample, the Apache resource depends on the MySQL resource. You can use theRemoteGroup agent to establish this dependency between these two resources.

With the RemoteGroup agent, you can monitor or manage a service group thatexists in a remote cluster.

Some points about configuring the RemoteGroup resource follow:

■ For each remote service group that you want to monitor or manage, you mustconfigure a corresponding RemoteGroup resource in the local cluster.

■ Multiple RemoteGroup resources in a local cluster can managecorrespondingmultiple remote service groups in different remote clusters.

■ You can include the RemoteGroup resource in any kind of resource or servicegroup dependency tree.

■ A combination of thestateof the local servicegroupand thestateof theremoteservice group determines the state of the RemoteGroup resource.

Symantec supports the RemoteGroup agent when:■ When it points to a global group

The RemoteGroup agent must then map the state of the global group in thelocal cluster.

■ When it is configured inside a local parallel service groupThe RemoteGroup resources on all cluster nodes monitor the same remoteservice group unless its attributes are localized.

■ When it is configured inside a local failover service group

For more information on thefunctionality of this agent refer to the VeritasCluster Server Administrator ’ s Guide .

DependencyAs a best practice, establish a RemoteGroup resource dependency on a NICresource. Symantec recommends that the RemoteGroup resource not be by itself in a service group.

Infrastructure and support agentsRemoteGroup agent



Agent functions

Brings the remote service group online. For more information:


Online

Takes the remote service group offline. For more information:


Offline

Monitors the state of the remote service group.

The true state of the remote service group is monitored only onthe online node in the local cluster. For more information:


Monitor

If theRemoteGroupresource faults, theCleanfunctiontakestheremote service group offline. For more information:


Clean

State definitions

Indicates that the remote service group is in an ONLINE state.

If the ReturnIntOffline attribute is not set to RemotePartial, then theremote service group is either in an ONLINE or PARTIAL state.

ONLINE

Indicatesthat theremote service group is inan OFFLINE or FAULTEDstate. The true state of the remote service group is monitored only onthe online node in the local cluster.

TheRemoteGroup resourcereturns intentionaloffline if theattributeReturnIntOffline is set to an appropriate value.

OFFLINE

Indicates that the RemoteGroup resource has unexpectedly goneoffline.

FAULTED

Indicates that a problem exists either with the configuration or theability of the RemoteGroup resource to determine the state of theremote service group.

UNKNOWN


226





The IP address or DNS name of a node in the remote cluster. The IPaddress can be either physical or virtual.

When configuring a virtual IP address of a remote cluster, do notconfigure the IP resource as a part of the remote service group.


Examples: "www.example.com" or "11.183.12.214"

IpAddress

This is a required attribute when the remote cluster listens on a portother than the default value of 14141.


Port

The name of the service group on the remote cluster that you wantthe RemoteGroup agent to monitor or manage.


Example: "DBGrp"

GroupName

You must set this attribute to either the VCS systemname or the ANYvalue.

■ ANYTheRemoteGroupresourcegoes online if theremote servicegroupis online on any node in the remote cluster.

■ VCSSysName Use the name of a VCS system ina remote cluster where you wantthe remote service group to be online when the RemoteGroupresource goes online. Use this to establish a one-to-one mappingbetween the nodes of the local and remote clusters.


Example: "vcssys1" or "ANY"

VCSSysName






Select only one of these values to determine the mode of operation of the RemoteGroup resource: MonitorOnly, OnlineOnly, or OnOff.

■ OnOff TheRemoteGroupresourcebrings theremote servicegroup onlineor takes it offline.When you set the VCSSysName attribute to ANY, the SysListattribute of the remote service group determines the node wherethe remote service group onlines.

■ MonitorOnlyThe RemoteGroup resource only monitors the state of the remoteservicegroup. TheRemoteGroup resourcecannot online or offlinethe remote service group.Make sure that you bring the remote service group online beforeyou online the RemoteGroup resource.

■ OnlineOnlyThe RemoteGroup resource only brings the remote service grouponline.TheRemoteGroup resourcecannot taketheremote servicegroup offline.When you set the VCSSysName attribute to ANY, the SysListattribute of the remote service group determines the node wherethe remote service group onlines.


ControlMode


228





This is the login user name for the remote cluster.

When you set the ControlMode attribute to OnOff or OnlineOnly, theUsername musthaveadministrative privileges for the remote servicegroup that you specify in the GroupName attribute.

When youuse theRemoteGroupWizard to enter your username data,you need to enter your username and the domain name in separatefields. For a cluster that has the Symantec Product AuthenticationService, you do not need to enter the domain name.

For a secure remote cluster:

■ Local Unix useruser@nodename — where the nodename is the name of the nodethat is specified in the IpAddress attribute. Do not set theDomainType attribute.

■ NIS or NIS+ useruser@domainName — where domainName is the name of the NISor NIS+ domain for the user. You must set the value of theDomainType attribute to either to nis or nisplus.


Example:

■ Fora cluster without theSymantecProduct AuthenticationService:"johnsmith"

■ For a secure remote cluster: "[email protected]"

Username

This is the password that corresponds to the user that you specify inthe Username attribute. You must encrypt the password with thevcsencrypt -agent command.

Note: Do notusethevcsencrypt utilitywhen entering passwords froma configuration wizard or the Cluster Manager (Java Console).


Password






For a secure remote cluster only, enter the domain typeinformation for the specified user.

For users who have the domain type unixpwd, you do not haveto set this attribute.

Type: string-scalar

Example: "nis", "nisplus"

DomainType

For a secure remote cluster only. If you need the RemoteGroupagent tocommunicate to a specific authentication broker, set thevalue of this attribute to the broker ’ s IP address.

Type: string-scalar

Example: "128.11.295.51"

BrokerIp

The port where the remote engine listens for requests.

This is an optional attribute, unlessthe remote cluster listens ona port other than the default value of 14141.


Default: 14141

Port

The maximum expected time in seconds that the remote service

group may take to offline. VCS calls the clean function for theRemoteGroup resource if theremote service group takesa longertime to offline than the time that you have specified for thisattribute.


Default: 0

OfflineWaitTime


230





Select one of the following values for RemoteGroup to returnIntentionalOffline:

■ RemotePartial — Indicates that the RemoteGroup resourcereturns an IntentionalOffline if the remote service group isin an ONLINE|PARTIAL state.

■ RemoteOffline — Indicates that the RemoteGroup resourcereturns an IntentionalOffline if the remote service group isin an OFFLINE state.

■ RemoteFaulted — Indicates that the RemoteGroup resourcereturns an IntentionalOffline if the remote service group isOFFLINE|FAULTED.

You can use these values in combinations with each other.

Youmust setthe IntentionalOfflineattribute of theRemoteGroupresource type to 1 for this attribute to work properly. For moreinformation about this attribute, see the Veritas Cluster Server Administrator ’ s Guide .


Default: ""

ReturnIntOffline

Defines the cluster node that performs the offline monitoring of the remote service group. This is an internal attribute. Do notmodify.

OfflineMonitoringNode

Table 6-7 Type-level attributes

DescriptionType levelattributes

In case of remote service groups that take a longer time to Online,Symantec recommends that youmodify the default OnlineWaitLimitand OnlineRetryLimit attributes.

See the Veritas Cluster Server Administrator ’ s Guide for moreinformation about these attributes.

OnlineRetryLimit

OnlineWaitLimit

If you expect the RemoteGroup agent to tolerate sudden offlines of the remote service group, then modify the ToleranceLimit attribute.

See the Veritas Cluster Server Administrator ’ s Guid e for moreinformation about these attributes.

ToleranceLimit

MonitorInterval




Testing agents


■ About the testing agents■ ElifNone agent

■ FileNone agent

■ FileOnOff agent

■ FileOnOnly agent

About the testing agentsUsethe testingagents to provide high availability forprogram support resources.These resources are useful for testing service groups.

ElifNone agentThe ElifNone agent monitors a file. It checks for the file ’ s absence.

You can use the ElifNone agent to test service group behavior. You can also useit as an impostor resource, where it takes the place of a resource for testing.

Dependencies for ElifNone agentNo dependencies exist for the ElifNone resource.

7Chapter



Figure 7-1 Sample service group that includes an ElifNone resource

Application

IP

NIC

Mount

Volume

DiskGroup

ElifNone

Agent function for ElifNone agentChecks for the specified file. If it exists, the resource faults. If it doesnot exist, the agent reports the resource as ONLINE.

Monitor

State definitions for ElifNone agent

Indicates that the file specified in the PathName attribute does notexist.

ONLINE

Indicates that the file specified in the PathName attribute exists.FAULTED

Indicates that the value of the PathName attribute does not containa file name.

UNKNOWN

Attributes for ElifNone agentTable 7-1 Required attribute


Specifiesthe complete pathname.Startswith a slash (/)precedingthefile name.


Example: "/tmp/file01"

PathName

Testing agentsElifNone agent

234



Resource type definition for ElifNone agenttype ElifNone (

static str ArgList[] = { PathName }static int OfflineMonitorInterval = 60static str Operations = Nonestr PathName

)

Sample configuration for ElifNone agentElifNone tmp_file01 (

PathName = "/tmp/file01")

Debug log levels for ElifNone agentThe ElifNone agent uses the following debug log levels:

DBG_4, DBG_5

FileNone agentMonitors a file, checks for the file ’ s existence.

You can use the FileNone agent to test service group behavior. You can also useit as an "impostor" resource, where it takes the place of a resource for testing.

Dependencies for FileNone agentNo dependencies exist for the FileNone resource.

Figure 7-2 Sample service group that includes an FileNone resource

Application

IP

NIC

Mount

Volume

DiskGroup

FileNone

Testing agentsFileNone agent



Agent functions for FileNone agent

Checksfor thespecified file. If it exists, theagentreports theresource

as ONLINE. If it does not exist, the resource faults.

Monitor

State definitions for FileNone agent

Indicates that the file specified in the PathName attribute exists.ONLINE


FAULTED


UNKNOWN

Attribute for FileNone agentTable 7-2 Required attribute



Type and dimension: string-scalarExample: "/tmp/file01"

PathName

Resource type definition for FileNone agenttype FileNone (

static str ArgList[] = { PathName }static int OfflineMonitorInterval = 60static str Operations = Nonestr PathName

)

Sample configuration for FileNone agentFileNone tmp_file01 (


Testing agentsFileNone agent

236



Debug log levels for FileNone agentThe FileNone agent uses the following debug log levels:

DBG_4, DBG_5

FileOnOff agentThe FileOnOff agent creates, removes, and monitors a file.


Dependencies for FileOnOff agentNo dependencies exist for the FileOnOff resource.

Figure 7-3 Sample service group that includes a FileOnOff resource

Application

IP

NIC

Mount

Volume

DiskGroup

FileOnOff

Agent functions for FileOnOff agent

Creates an empty file with the specified name if the file does notalready exist.

Online

Removes the specified file.Offline

Checks for the specified file. If it exists, the agent reports as ONLINE.If it does not exist, the agent reports as OFFLINE.

Monitor

Removes the specified file forcibly when necessary.Clean

Testing agentsFileOnOff agent



State definitions for FileOnOff agent



OFFLINE

Indicates that the file specified in the PathName attribute has beenremoved out of VCS control.

FAULTED


UNKNOWN

Attribute for FileOnOff agentTable 7-3 Required attribute





PathName

Resource type definition for FileOnOff agenttype FileOnOff (

static str ArgList[] = { PathName }str PathName

)

Sample configuration for FileOnOff agentFileOnOff tmp_file01 (


Debug log levels for FileOnOff agentThe FileOnOff agent uses the following debug log levels:

DBG_1, DBG_4, DBG_5

Testing agentsFileOnOff agent

238



FileOnOnly agentThe FileOnOnly agent creates and monitors a file.


Dependencies for FileOnOnly agentNo dependencies exist for the FileOnOnly resource.

Figure 7-4 Sample service group that includes a FileOnOnly resource

Application

IP

NIC

Mount

Volume

DiskGroup

FileOnOnly

Agent functions for FileOnOnly agentCreates an empty file with the specified name, unless one alreadyexists.

Online

Checks for the specified file. If it exists, the agent reports as ONLINE.If it does not exist, the resource faults.

Monitor


Clean

State definitions for FileOnOnly agentThe state definitions for this agent follow:


Indicates that the file specified in the PathName attribute does notexist and VCS has not attempted to bring the resource online.

OFFLINE

Testing agentsFileOnOnly agent



Indicates that the file specified in the PathName attribute has beenremoved out of VCS control.

FAULTED


UNKNOWN

Attribute for FileOnOnly agentTable 7-4 Required attributes


Specifiesthe complete pathname.Startswith a slash (/)precedingthe

file name.Type and dimension: string-scalar


PathName

Resource type definition for FileOnOnly agenttype FileOnOnly (

static str ArgList[] = { PathName }static str Operations = OnOnlystr PathName

)

Sample configuration for FileOnOnly agentFileOnOnly tmp_file02 (PathName = "/tmp/file02")

Debug log levels for FileOnOnly agentThe FileOnOnly agent uses the following debug log levels:

DBG_1, DBG_4, DBG_5

Testing agentsFileOnOnly agent

240



Replication agents


■ About the replication agents■ RVG agent

■ RVGPrimary agent

■ RVGSnapshot

■ RVGShared agent

■ RVGLogowner agent

■ RVGSharedPri agent

■ VFRJob agent

About the replication agentsUse the replication agents to provide high availability for VVR resources.

Refer to the Veritas Storage Foundation andHighAvailabilitySolutions Replication Administrator's Guide for information on configuring the Replication agents forhigh availability.

RVG agentBrings the RVG online, monitors read and write access to the RVG, and takes theRVGoffline. This is a failover resource.TheRVGagent enables replication betweenclusters. It manages the Primary VVR node in one cluster and the Secondary VVRnode in another cluster. Each node can be failed over in its respective cluster. Inthis way, replication is made highly available.

8Chapter



The RVG agent manages the state of the RVG during local failovers. TheRVGPrimary agent manages the role of the RVG during a wide area failover.

Usinga VCSglobal clusterenables you to fail over thePrimary role from a PrimaryVVR node to a Secondary VVR node.The RVG agent includes the following key features:

■ Removes potential singlepoints of failure by enabling Primary andSecondaryVVR nodes to be clustered.

■ Enablesyouto bring a service group online to start VCS-managed applicationsthat use VVR.

■ Continues replication after a node in a cluster fails without losing updates.

■ Ensures that VVR can be added to any VCS cluster by including the RVG

resource type definitions.An example configuration file for this agent that can be used as a guide whencreating your configuration is located at:

/etc/VRTSvcs/conf/sample_vvr/RVG

DependenciesThe RVG resource represents the RVG (Replicated Volume Group) in the RDS(Replicated Data Set). TheRVGresource is dependent on theDiskGroup resource.TheRVG resource is also dependenton theIP resources that it uses for replication.

Refer to the Veritas Cluster Server Administrator's Guide for more informationon dependencies.

Figure 8-1 Sample service group for an RVG resource

NIC

IPDiskGroup

Mount

RVG

Replication agentsRVG agent

242



Agent functionsThe RVG agent has the following agent functions:

Verifies whether the DiskGroup agent has recovered the RVG. If not,recovers andstarts thedata volumes andthe Storage Replicator Log(SRL),recoversthe RVG, recoversallRLINKs in the RVG, andthenstarts theRVG.

Online

Stops the RVG.Offline

Monitors the state of the RVG using the vxprint command.

The RVG resource monitors an RVG for local access only. It does notmonitor replication.

Monitor

Stops the RVG.Clean

The info entry point displays information about the replication status of a RDS.

Info

State definitionsThe RVG agent has the following state definitions:

Indicates that the RVG is in ENABLED /ACTIVE state.ONLINE

Indicates that the RVG is in DISABLED /CLEANstate.OFFLINE

The RVG resource fails if the RVG is not in the ENABLED/ACTIVE state.FAULTED


DescriptionRequiredattributes

The name of the RVG being monitored.


Example: "hr_rvg"

RVG

The disk group that this RVG is associated with.


Example: "hrbg"

DiskGroup

Replication agentsRVG agent



RVGPrimary agentThe RVGPrimary agent enables migration and takeover of a VVR Replicated

Volume Group (RVG) in a VCS environment. Bringing a resource of typeRVGPrimary online causes the RVG on the local host to become a primary.

The agent is useful when hosts in both the primary and secondary side areclustered, in particular a VCS replicated data cluster or a VCS global cluster, tocompletely automate theavailabilityofwritable replicateddisks to a VCS-managedapplication.

The RVGPrimary agent includes the following features:

■ Removes the manual steps of migrating a VVR primary and secondary roleswhen failing over applications across a wide area.

■ Minimizes the need for resynchronizing replicated volumes by attempting amigration before attempting a hard takeover.

■ Waits for the two sides of a replicated data set to become completelysynchronized before migrating roles.

■ Supports an automatic fast failback resynchronization of a downed primaryif it later returns after a takeover.

■ Allows you to distinguish the Primary site after network failure or disaster

■ Supports the ability to choose the Primary site after a site failure or networkdisruption is corrected.

■ Aftera successful migrationor takeover of a SecondaryRVG, theRVGPrimaryagent ensures to automatically start the replication from the new Primary toany additional Secondary(s) that exists in the RDS.

■ Before a takeover, the RVGPrimary agent synchronizes the Secondary sitewith any bunker associated with the Primary site, when the Primary site isnot available.

Refer to the Veritas Storage Foundation andHighAvailabilitySolutions Replication Administrator's Guide for information on configuring the Replication agents forhigh availability.

A sample configuration file for this agent that you can use as a guide to createthe configuration is located at /etc/VRTSvcs/conf/sample_vvr/RVGPrimary.

DependenciesYou usually use the RVGPrimary agent in conjunction with the RVG agent in twogroups with an online local hard group dependency. The parent group containsthe resources that manage the actual application and file systems and as the

Replication agentsRVGPrimary agent



RVGPrimary resource. The child group contains the resources managing thestorage infrastructure, which include the RVG and DiskGroup type resources.

Refer to the Veritas Storage Foundation andHighAvailabilitySolutions Replication

Administrator's Guide for information about the setup of a VVR environmentusing the RVGPrimary agent.

Figure 8-2 Sample service group for an RVGPrimary resource

online local hard dependency

Application service group that is online on the Primary

Oracle

RVGPrimary

MountMount Mount

NIC

IP

RVG

DiskGroup IP

NIC

Replication service group that is online at both Primary and Secondary

Agent functionsThe RVGPrimary agent has the following agent functions:

Determines thecurrent role of theRVG. If therole isSecondaryit attemptsa migration. It waits for anyoutstanding writes from the original Primary.If the original Primary is down, it attempts a takeover. You can configurethe RVGPrimary agent so that, before a takeover, the agent synchronizestheSecondary site with anybunkerassociated with thePrimary site, whenthe Primary site is not available. If the RVG is a Primary, it performs noactions and goes online.

Online

Performs no actions.Offline


246





Indicates whether the agent should attempt to automaticallyperform a fast-failback resynchronization of the original Primaryafter a takeover and after the original Primary returns.

You can use the following values for this attribute:

■ 0– instructs the agent to not attempt to perform a fast-failbackresynchronization of theoriginal Primary after a takeover andafter the original Primary returns.

■ 1– instructs the agent to attempt to automatically perform afast-failback resynchronization of the original Primary after atakeover and after the original Primary returns.

■ 2– instructstheagenttousethe primary-electfeature.The agentdoes not attempt to perform a fast-failback resynchronizationof the original Primary after a takeover and after the originalPrimary returns. The RVGPrimary agent also createsspace-optimizedsnapshots for all thedata volumes in theRVGresource.If you set the AutoResync attribute to 2 (to enable theprimary-elect feature) the value of the BunkerSyncTimeOutattribute must be zero to disable the automated bunker replayfeature. You cannot use the automated bunker replay featureand the primary-elect feature in the same environment.

AutoTakeover and AutoResync are mutually exclusive attributes.When AutoTakeover=0, theprimary-elect featureisnotapplicable;therefore, it is not supported.


AutoResync


248



"RvgResourceName:DiskGroup", AutoTakeover, AutoResync,BunkerSyncTimeOut, BunkerSyncElapsedTime }str RvgResourceName

int AutoTakeover = 1int AutoResync = 0str BunkerSyncTimeOutint BunkerSyncElapsedTime = 0

)

Sample configurationsRVGPrimary rvg-pri (

RvgResourceName = rvgRes)

RVGSnapshotFor a fire drill, creates and destroys a transactionally consistent space-optimizedsnapshotofallvolumes in a VVRsecondary replicated data set. TheRVGSnapshotagent takes space-optimized snapshotson a secondaryRVG. Thesesnapshotscanbe mounted and written to without affecting the actual replicated data, whichmeans that the space-optimized snapshot can be an effective tool for schedulinga “ fire drill ” to confirm that a wide-area failover is possible. By combining thisagent with theVCSMount agent, theCFSMount agent,andVCSagents that managethe application being replicated, you can create a special fire drill service group.You can bring this service group online and take it offline at regularly scheduledintervals to verify that the disaster recovery environment is robust.

In addition to theagent itself, a text-based wizard/opt/VRTSvcs/bin/fdsetup thatprepares the VVR and VCS infrastructure for a fire drill and a script /opt/VRTSvcs/bin/fdsched that runs the fire drill and consolidates the resultsare also included.

Complete details are in the Veritas Cluster Server Administrator's Guide .

The RVGSnapshot agent includes the following key features:

■ Automates the process of creating a space-optimized snapshot on a VVRsecondary that can be mounted to simulate a wide-area failover withoutaffecting the production application.

■ Includes a wizard to effectively set up and schedule fire drills that arecompletely managed by VCS.

Replication agentsRVGSnapshot





The name of the VCS RVG-type resource that manages the RVGthat will be snapshot by this agent.


RvgResourceName

Name of the cache object that is required for a space-optimizedsnapshot; the fdsetup wizard will create one if it does not exist


CacheObj

Token putbefore thename of theactualvolumewhen creating the

snapshotted volumes.


Prefix



A flag to indicate whether to destroy the snapshotupon takingtheresources offline. For a fire drill, the snapshot should be deletedto reduce any performance impact of leaving the snapshot for along period of time; however, if there is interest in keeping the

data, then this value should be set to 0. The default is 1 (true).Type and dimension: integer-scalar

Default: 1

DestroyOnOffline

The fire drill schedule updates this attributewith thesystem nameand the path to a file containing the output of the last completefire drill for the group containing an RVGSnapshot resource.


FDFile

Resource type definitionsThe resource type definition for the RVGSnapshot agent follows.

type RVGSnapshot (static keylist RegList = { Prefix }static int NumThreads = 1static str ArgList[] = { RvgResourceName, CacheObj, Prefix,

Replication agentsRVGSnapshot



DestroyOnOffline }str RvgResourceNamestr CacheObj

str Prefixboolean DestroyOnOffline = 1temp str FDFiletemp str VCSResLock

)

Sample configurationsRVGSnapshot rvg-sos (

RvgResourceName = ApplicationRVGCacheObj = cacheobj

Prefix = snap)

RVGShared agentMonitors the RVG in a shared environment. This is a parallel resource. TheRVGShared agent enables you to configure parallel applications to use an RVGin a cluster. The RVGShared agent monitors the RVG in a shared disk groupenvironment. The RVGShared agent must be configured as a parallel group inVCS. Typically, the RVGShared resource is online or offline at the same time onall the nodes in the VCS cluster. An example configuration file for this agent thatcan be used as a guide when creating your configuration is located at /etc/VRTSvcs/conf/sample_vvr/RVGLogowner.

DependenciesTheRVGSharedresource represents theRVGof theRDS. TheRVGSharedresourceis dependent on the CVMVolDg resource.

The RVGShared resource must be configured in a parallel group.

Refer to the Veritas Storage Foundation andHighAvailabilitySolutions Replication Administrator's Guide for information on configuring parallel applications forhighly availability.

Refer to the Veritas Cluster Server Administrator's Guide for more informationon dependencies.

Replication agentsRVGShared agent

254



Figure 8-4 Sample service group for an RVGShared resource

RVGShared

CVMVolDg

racdata_rvg

RVGShared Group (Parallel)

racdata_voldg

Note: Donot add any volumes that are part of theRVG in the CVMVolumeattributeof the CVMVolDg resource. The volumes in the RVG are managed by theRVGShared resource.

Agent functionsThe RVGShared agent has the following agent functions:

Verifies whether theRVGis started. If theRVGis notstarted, recovers andstarts the RVG.

Online

No action.Offline

Displays the state as ONLINE if the RVG is started. Displays the state asOFFLINE if the RVG is not started.

Monitor

No action.Clean

The info entry point displays information about the replication status of a RDS.

Info

State definitionsThe RVGShared agent has the following state definitions:

Indicates that the RVG is in the ENABLED /ACTIVE state.ONLINE

Indicates that the RVG is not in the ENABLED /ACTIVE state or that theadministrator has invoked the offline entry point.

OFFLINE

Replication agentsRVGShared agent







RVG

The shared-disk group with which this RVG is associated.


DiskGroup

Resource type definitionsThe RVGShared resource type definition follows.

type RVGShared (static int NumThreads = 1static str ArgList[] = { RVG, DiskGroup }str RVGstr DiskGroup

)

Sample configurationsRVGShared racdata_rvg (RVG = rac1_rvgDiskGroup = oradatadg)

RVGLogowner agentAssigns and unassigns a nodeas the logowner in the CVM cluster; this is a failoverresource. The RVGLogowner agent assigns or unassigns a node as a logowner in

the cluster. To replicate data, VVR requires network connectivity between thePrimary and the Secondary. In a shared disk group environment, only one node,that is, the logowner, can replicate data to the Secondary.

For replication to be highly available, the logowner must be highly available. Tomakethe logownerhighlyavailable, theRVGLogownerresourcemustbeconfiguredas a resource in a failover group. Also, a virtual IPmust be set up on the logownerto enable replication and failover of the logowner from one node to another in acluster. The virtual IP must be configured as an IP resource.

Replication agentsRVGLogowner agent

256



Formore information about the logowner,seethe Veritas StorageFoundationand High Availability Solutions Replication Administrator's Guide . An exampleconfiguration file for this agent that can be used as a guide when creating your

configuration, is located at /etc/VRTSvcs/conf/sample_vvr/RVGLogowner.

DependenciesThe RVGLogowner resource represents the logowner for RVG in the cluster. TheRVGLogowner resource is dependent on the IP resourcethat it uses for replication.

The RVGLogowner resource must be configured in a failover group. TheRVGLogownergroup is usedin conjunctionwiththeRVGSharedPriandRVGSharedagents in separate groups, with the appropriate service group dependencies.

For more information on dependencies, refer to the Veritas Cluster Server

Administrator's Guide

Figure 8-5 Sample service group for an RVGLogowner resource

RVGLogowner

IP

NIC

rvg_logowner

logowner_ip

nic

Logowner Group (Failover)

Agent functionsThe RVGLogowner agent has the following agent functions:

Assigns the logowner on the node.Online

Unassigns the logowner on the node.Offline

Returns ONLINE if the node is the logowner and the RVG is inENABLED/ACTIVE state. Returns OFFLINE if the node is the logowner andthe state is not ENABLED/ACTIVE , or if the node is not the logowner(regardless of the state). The RVG for which the logowner is monitoredmust be configured as the RVGShared resource type.

Monitor

Unassigns the logowner on the node.Clean




State definitionsThe RVGLogowner agent has the following state definitions:

Indicates that the node is the logowner for the RVG in the cluster.ONLINE

Indicates that the node is not the logowner for the RVG in the cluster.OFFLINE





Example: "hr_rvg"

RVG

The disk group with which this RVG is associated.


Example: "hrbg"

DiskGroup


DescriptionBunker attributesFor internal use only, do not modify. The name of the bunker diskgroup.


Example: "hr_bdg"

StorageDG

For internal use only, donot modify. The name of the bunker RVG.


Example: "hr_brvg"

StorageRVG

For internal use only, do not modify. A space-separated list of thehostids of each node in the bunker cluster.

Type and dimension: string-keylist

Example: "bunker_host"

StorageHostIds


258



Resource type definitionsThe RVGLogowner resource type definition follows.

type RVGLogowner (static int NumThreads = 1static str ArgList[] = { RVG, DiskGroup }static int OnlineRetryLimit = 5str RVGstr DiskGroupstr StorageRVGstr StorageDGstr StorageHostIds

)

RVGLogowner agent notesThe RVGLogowner agent has the following notes:

CVM master node needs to assume the logowner role for VCSmanaged VVR resourcesIf you use VCS to manage RVGLogowner resources in an SFCFSHA environmentor an SF Oracle RAC environment, Symantec recommends that you perform thefollowingprocedures. Theseprocedures ensure that theCVMmasternode always

assumes the logowner role. Not performing these procedures can result inunexpected issues that are due to a CVM slave node that assumes the logownerrole.

For a service group that contains an RVGLogowner resource, change the value of its PreOnline trigger to 1 to enable it.

To enable the PreOnline trigger from the command line on a service group that hasan RVGLogowner resource

On each node in the cluster, perform the following command:

# hagrp -modify RVGLogowner_resource_sg PreOnline 1 -sys system

Where RVGLogowner_resource_sg is the service group that contains theRVGLogowner resource. The system is the name of the node where you wantto enable the trigger.

On each node in the cluster, merge the preonline_vvr trigger into the defaulttriggers directory.




To merge the preonline_vvr trigger

On each node in the cluster, merge the preonline_vvr trigger to the /opt/VRTSvcs/bin/triggers directory.

# cp /opt/VRTSvcs/bin/sample_triggers/VRTSvcs/preonline_vvr \/opt/VRTSvcs/bin/triggers

Refer to the sample configurations directory for samples of how to enablethese triggers (/opt/VRTSvcs/bin/sample_triggers/VRTSvcs.)

Sample configurationsRVGLogowner vvr_rvglogowner (

RVG = app_rvgDiskGroup = vvrdg)

RVGSharedPri agentAttempts to migrate or takeover a Secondary to a Primary when a parallel servicegroup fails over. The RVGSharedPri agent enables migration and takeover of aVVR replicated data set in parallel groups in a VCS environment. Bringing aresource of type RVGSharedPrionlinecauses theRVGon the local host to becomea primary if it is not already. The agent is useful when hosts in both the primaryandsecondaryside areclusteredusinga VCSglobalcluster, to completelyautomatethe availability of writable replicated disks to an application managed by VCS.

Youcannotusetheprimary-elect featurewiththisagent.Fora detailed descriptionof theprimary-elect feature, see Veritas Storage FoundationandHighAvailability Solutions Replication Administrator's Guide .

The RVGSharedPri agent includes the following key features:

■ Removes manual steps of migrating a VVRprimary andsecondary roles whenfailing over applications across a wide area.

■ Minimizes the need for resynchronizing replicated volumes by attempting amigration before attempting a hard takeover.

■ Waits for the two sides of a replicated data set to become completelysynchronized before migrating roles.

■ Supports an automatic fast failback resynchronization of a downed primaryif it later returns after a takeover.

Sample configuration files are located in the /etc/VRTSvcs/conf/sample_rac/directoryand include CVRin thefilename.These samplefiles are installedas part

Replication agentsRVGSharedPri agent

260



of the VRTSdbac RPM, and can be used as a guide when creating yourconfiguration.

DependenciesThe RVGSharedPri agent is used in conjunction with the RVGShared andRVGLogowner agents in separate groups, with the appropriate service groupdependencies.

Refer to the Veritas Storage Foundation andHighAvailabilitySolutions Replication Administrator's Guide for information on configuring parallel applications forhighly availablity.

The RVGSharedPri agent must be configured in a parallel service group. Theapplication service group contains the resourcesmanaging the actual application

and file systems as well as the RVGSharedPri agent.

Figure 8-6 Sample service group for an RVGSharedPri resource

Oracle

CFSMount

ora_db1

ora_vvr_shpri

RVGSharedPri

Application Group (Parallel)

Agent functionsThe RVGSharedPri agent has the following agent functions:

Determines the current role of the RVG; if Secondary, attempt a migrate,waitingforanyoutstandingwrites fromthe original Primary; if theoriginalPrimary is down attempt a takeover; if the RVG is a Primary, perform no

actions and go online

Online

Performs no actions.Offline

Performs no actions; monitoring of the actual RVG is done by theRVGShared agent.

Monitor

Performs no actions.Clean




This is an action entry point.

It resynchronizes the originalPrimary with the new Primary that has takenoverwithfast-failback, after the original Primary had become unavailable.

This needs to be executed when the original Primary becomes availableand starts acting as a Secondary.

fbsync

This is an action entry point.

It resynchronizes the Secondaries with the Primary using DCM.

resync

State definitionsThe RVGSharedPri agent has the following state definitions:

Monitoring of the actual RVG is done by the RVGShared agent; accidentalmigration of a VVR Primary outside of VCS would cause other resourcesto fault immediately, such asMount,so nospecial monitoring by this agentis necessary.

FAULTED



The name of the RVGShared resource type that this agent willpromote, that is, the name RVG resource type which has beenconfigured using the RVGShared agent.


RvgResourceName

A flag to indicate whether the agent should perform a takeover ononline if the original Primary is down.


Default: 1

AutoTakeover

A flag to indicatewhether theagentshould attempt toautomaticallyperform a fast-failback resynchronization of the original Primaryafter a takeover and after the original Primary returns.


Default: 0

AutoResync

This attribute is reserved for internal use by VCS.


VCSResLock


262



Resource type definitionsThe RVGSharedPri resource type definition follows.

type RVGSharedPri (static keylist SupportedActions = { fbsync, resync }static int NumThreads = 1static int OnlineRetryLimit = 1static str ArgList[] = { RvgResourceName, "RvgResourceName:RVG","RvgResourceName:DiskGroup", AutoTakeover, AutoResync }str RvgResourceNameint AutoTakeover = 1int AutoResync = 0temp str VCSResLock

)

Sample configurationsRVGSharedPri ora_vvr_shpri (RvgResourceName = racdata_rvgOnlineRetryLimit = 0)

VFRJob agentVeritas File Replicator Job (VFRJob) agent provides high availability for VeritasFile SystemReplicator Job (VFR Job).VFRJob schedules replication of file systemsfrom a source system to a target system. The agent makes the VFR Job highlyavailable on a source system. VFR Job supports replication of VxFS and CFS typefile systems.

Refer to Veritas Storage Foundation ™ and HighAvailability SolutionsReplicationAdministrator's Guide for more details of VFR.

OverviewThe VFRJob Agent starts scheduling of VFR Job, monitors VFR Job status, andstops scheduling of VFR Job.

Use the VFRJob agent to make the replicator job highly available on a sourcesystem if thesource system faults. TheVFRJob type resource is a failover resourceand provides high availability (HA) for VFR Job. It monitors the VFR Job on thesource system. The source system is where the file system is mounted and the filesystem is replicated from the source system to the target system. The target

Replication agentsVFRJob agent



system, where the file system is replicated to, must be outside of the cluster of the source system. Target system can be part of a different cluster.

If thesystem performingfile system replication (systemthat hosts thefile system)faults, the file system fails over to another system in the cluster. And the VFRJobresource also fails over to that system. Thus VFRJob agent makes the VFR Jobhighly available.

VFRJob depends on vxfstaskd daemon and vxsfsrepld daemon. The vxfstaskddaemon schedules the replicator job while the vxfsrepld daemon functions as thereplicator daemon. Both the daemons should be running on source and targetsystems.

Refer to the sample configuration that shows how VCS provides high availabilityfor the scheduler daemon and the replication daemon on the source as well as thetarget system..

Dependencies for VFRJob agentThe VFRJob resource represents the VFR Job. The resource is dependent on thevxfstaskd daemon and vxfsrepld daemon. The vxfstaskd daemon is a schedulerdaemon and the vxfsrepld daemon is a replicator daemon.

Refer to the Veritas Cluster Server Administrator's Guide for more informationon resource and group dependencies.

See “ Group dependency for VFRJob agent in Custer File Systems or VxFS File

Systems”

on page 264.See “ Group dependency for VFRJob agent in VxFS File Systems ” on page 265.

Group dependency for VFRJob agent in Custer File Systems orVxFS File SystemsFor Cluster File System (CFS) or VxFS, configure the VFRJob service group andits dependency as follows:


264



Figure 8-7 with VFRJob resource in a separate service group than the servicegroup representing VxFS or CFS file system resource

When the VFRJob agent fails for a VxFS file system and you want to failover theVxFS file system, you must configure thegroup dependency for VFRJob with VxFS

file systems.See “ Group dependency for VFRJob agent in VxFS File Systems ” on page 265.

Group dependency for VFRJob agent in VxFS File SystemsConfigure the VFRJob and VxFS file system resources as described in this sectionif you want to failover the file system when the VFRJob faults.

Consider a VxFS file system and VFRJob resource configured in separate servicegroups. If the VFRJob faults, the VFRJob does not failover to another system. TheVFRJob does not failover because the file system is still mounted on the system

where the VFRJob faulted.■ Include the VFRJob resource in the service group that represents the file

system. Add resource dependency for the VFRJob resource, that is, VFRJobrequires Mount resource where the file system is mounted.

■ Add group dependency between the service group that contains the VFRJobandtheservice group that contains thedaemonresources. Dependency betweenthe groups must be set to online local firm.




Figure 8-8 VFRJob resource and VxFS file system resource configured in thesame service group

High availability of scheduler and replicator daemonsYou can configure the vxfstaskd and vxfsrepld daemons which are representedas application type resources in separate service groups of the type parallel.

See “ High availability of VFR daemons ” on page 268.

Agent functions for VFRJob agentTable 8-10 Agent functions

DescriptionAgent function

Starts scheduling of VFRJob.Online

Stops scheduling of VFRJob.Offline


266



Table 8-10 Agent functions (continued)

Monitors the state of the VFRJob using vfradmin command.The agent does not monitor replication.

Monitor

Stops scheduling of VFRJob.Clean

State definitions for VFRJob agentTable 8-11 State definitions

DescriptionState definition

Indicates Job is running or scheduled to run.Online

Indicates that the VFRJob is stopped or is not configured.Offline

Indicates VFRJob that was running earlier but is now offlineoutside VCS control.

Faulted

Attributes for VFRJob AgentTable 8-12 Required attributes

DescriptionAttribute

Name for the VFRJob.Type and dimension: String Scalar

JobName

IP address of the source system from where the file system isto be replicated.

Type and dimension: String Scalar

SrcAddress

Mount point on the source system from where the file systemis to be replicated.

Type and dimension: String Scalar

SrcMountPoint

Resource type definitions for VFRJob agentThe VFRJob agent resource type definition follows:

type VFRJob (static str ArgList[] = { JobName, SrcMountPoint, SrcAddress }str JobName




str SrcMountPointstr SrcAddress

)

High availability of VFR daemonsYou can make the vxfstaskd daemon and the vxfsrepld daemon highly availableby configuring these daemons as application type resources.

The following configurations are supported:

Configuration of VFRJob service groups on the source systemConsider the following service groups on a source system from where the filesystem is replicated to the target system:

■ Service group FSx_Grp represents a file system (either CVM/CFS or VM/FS)

■ Service group VFRJob_Grp represents the VFRJob

■ Service group VFRSourceInfra_Grp represents vxfstaskd and vxfsreplddaemons

The group dependencies can be defined as follows:

■ VFRJob_Grp requires FSx_Grp online local firm

■ VFRJob_Grp requires VFRSrcInfra_Grp online local firm

For more information on resource and group dependencies, refer to the VeritasCluster Server Administrator's Guide .

Sample configuration of VFRJob agent on source systemSample configuration of VFRJob on source system including resource and groupdependencies.

group VFRJob_Grp (SystemList = { sys1= 0, sys2= 1 })

VFRJob vfr_job (JobName = vfr_job_fs1SrcMountPoint = "/dgsfs1"SrcAddress = " 10.198.95.191"RestartLimit = 1)


268



■ Service group VFRJob_Grp2 represents theVFRjobgroup on the target system.It contains the virtual IP that is used for VFR replication on the target systemand VFRJob resource which is disabled (Enabled = 0 is set).

Note that the VFRJob resource always needs to be disabled on target system.Also, the disabled VFRJob resource on the target system is enabled only whendirection of replication needs to be changed for VFR.For more details, See “ Changing file replication direction ” on page 271.

■ Define the group dependencies as follows:

■ VFRJob_Grp2 requires FSx_Grp2 in an online local firm dependency.

■ VFRJob_Grp2 requires VFRTargetInfra_Grp in an online local firmdependency.

For more information on resource and group dependencies, refer to the VeritasCluster Server Administrator's Guide .

Sample configuration of VFRJob agent on target systemSample configuration for a target system (under VCS control) to provide highavailability for VFR replicator daemon and IP address used by replicator daemonon target systems.

group VFRJob_Grp2 (SystemList = {sysx= 0, sysy = 1}AutoStartList = {sysx= 0, sysy = 1})

VFRJob vfr_job2 (Enabled = 0JobName = vfr_job_fs2SrcMountPoint = "/dgsfs2"SrcAddress = " 10.198.95.192"RestartLimit = 1

)

IP vip_vfr2 (Device = eth0Address = "10.198.95.192"NetMask = "255.255.255.0")


270



vfr_job2 requires vip_vfr2requires group VFRTargetInfra_Grp online local firmrequires group FSx_Grp2 online local firm

Changing file replication directionIf a target file system is written to during a disaster, and then becomes theprimarysite, the replication direction can be changed so that the old source file systemcan be made the new target file system.

If both the source and target systems for VFRJob are under VCS control, performthe following steps to change the file replication direction for VFR.

Note: The vfr_job resource represents the VFRJob running on current sourcesystem while vfr_job2 resource represents the VFRJob resource that is disabled(Enabled 0) on the target system.

On the source system perform the following steps

1 Offline the VFRJob resource on source system

# hares -offline vfr_job -sys source_system ,

where vfr_job is theVFRJob resource andsource_system is the systemwhereVFRJob is online.

2 After the VFRJob goes offline, disable the VFRJob resource.

# haconf -makerw

# hares -modify vfr_job Enabled 0

# haconf -dump -makero

3 Perform the steps to change the direction of replication for VFR on both thesource system and target system.

For more details, refer to Veritas Storage Foundation ™ and High Availability Solutions Replication Administrator's Guide .

On the new source system perform the following steps:

After the file replication direction change, the earlier target system becomes thenew source system.




1 Enable VFRJob

# haconf -makerw

# hares -modify vfr_job2 Enabled 1# haconf -dump -makero

2 Probe the VFRJob on the new source system

# hares -probe vfr_job2 -sys new_source_system , wherenew_source_system is system where the virtual IP for VFRJob2 , that is,vfr_vip2 is online.

3 Bring the VFRJob online.

# hares -online vfr_job2 -sys new_source_system

Notes for the VFRJob agentThe notes for VFRJob agent are as follows:

■ Symantec recommends using Virtual IP when settingup the VFRJobresource.

Using a different port number for the replication daemonTo run the replication daemon on a different port number perform the followingsteps:

1 Offline the VFRJob resource.2 Offline the vfr_repld resource by running the hares -offline command.

3 Modify the vfr_repld resource definition for StartProgram andMonitorProcesses with the new port number.

4 Online vfr_repld resource by running the hares -online command

5 Online the VFRJob resource by running the hares -online command.

6 On the target system skip Step1 and Step5 as VFRJob is disabled.


272



Symbols802.1Q trunking 80

Aabout

Network agents 79Samba agents 146

agentmodifying 18

agent functions 143Apache HTTP server agent 162Application agent 177CoordPoint agent 189DiskGroup agent 22DiskGroupSnap agent 34DiskReservation agent 47DNS agent 115ElifNone agent 234FileNone agent 236FileOnOff agent 237FileOnOnly agent 239IP agent 81IPMultiNIC agent 94LVMLogicalVolume agent 57LVMVolumeGroup agent 60Mount agent 65MultiNICA agent 102

Linux 102NetBIOS agent 156NFS agent 132

Linux 132NFSRestart agent 137

Linux 137NIC agent 88NotifierMngr agent 214Phantom agent 224Process agent 204ProcessOnOnly agent 208Proxy agent 220RemoteGroup agent 226SambaServer agent 149

agent functions (continued)SambaShare agent 153Share agent 143Volume agent 52VolumeSet agent 54

agents. See RVG agentApache HTTP server 161Application 176

CoordPoint 187DiskGroup 21DiskGroupSnap 32DiskReservation 47DNS 114ElifNone 233FileNone 235FileOnOff 237FileOnOnly 239IP 80IPMultiNIC 93LVMLogicalVolume 56

LVMVolumeGroup 59Mount 65MultiNICA 100

Linux 100NetBIOS 155NFS 131NFSRestart 136NIC 86NotifierMngr 213Phantom 223Process 203ProcessOnOnly 207

Proxy 220RemoteGroup 225RVGLogowner. See RVGLogowner agentRVGPrimary. See RVGPrimary agentRVGShared. See RVGShared agentRVGSharedPri. See RVGSharedPri agentRVGSnapshot. See RVGSnapshot agentSambaServer 149SambaShare 153Share 142

Index



agents (continued)VFRJob 263Volume 51

Apache HTTP server agentagent functions 162attributes 164description 161detecting application failure 169sample configuration 171state definitions 163

Application agentagent functions 177attributes 180description 176high availability fire drill 176Linux attributes 180resource type definition 184sample configurations 186state definitions 179

association dimension 18attribute data types 18attributes

Application agent 180Linux 180

CoordPoint agent 190DiskGroup agent 24

Linux 24DiskGroupSnap agent 34DiskReservation agent 48

Linux 48DNS agent 117ElifNone agent 234FileNone agent 236FileOnOff agent 238FileOnOnly agent 240IP agent 82

Linux 82IPMultiNIC agent 95

Linux 95LVMLogicalVolume agent 58

Linux 58LVMVolumeGroup agent 61

Linux 61modifying 18Mount agent 68

Linux 68MultiNICA agent 103

Linux 103

attributes (continued)NFS agent 133

Linux 133NFSRestart agent 139

Linux 139NIC agent 89

Linux 89NotifierMngr agent 215

Linux 215Process agent 205

Linux 205ProcessOnOnly 208ProcessOnOnly agent

Linux 208Proxy agent 221RemoteGroup agent 227SambaServer agent 150Share agent 144

Linux 144Volume agent 53VolumeSet agent 55

AutoResync attributeRVGPrimary agent 262

Bbonded network interfaces 87boolean data types 18

CCluster Manager (Java Console)

modifying attributes 18CNAME record 126configuration files

main.cf 224modifying 18types.cf 18

CoordPoint agentagent functions 189attributes 190description 187resource type definition 190sample configurations 192state definitions 189

Ddata type

boolean 18string 18

Index274



data typesinteger 18

dependency graphsRVGLogowner agent 257RVGPrimary agent 245RVGShared agent 254RVGSharedPri agent 261

descriptionresources 18

dimensionskeylist 18scalar 18vector 18

DiskGroup agentagent functions 22attributes 24description 21high availability fire drill 30Linux attributes 24resource type definition 30sample configurations 32state definitions 24

DiskGroupSnap agentagent functions 34attributes 34description 32resource type definition 40sample configurations 40state definitions 34

DiskReservation agentagent functions 47attributes 48description 47Linux attributes 48resource type definition 49sample configurations 50state definitions 47

DNS agent 116agent functions 115attributes 117description 114resource type definition 123sample web server configuration 126

EElifNone agent

agent functions 234attributes 234description 233

ElifNone agent (continued)resource type definition 235sample configuration 235state definitions 234

Ffailover group

RVGLogowner agent 256fast failback

AutoResync attribute of RVGPrimary 262fast failback resynchronization

RVGPrimary 245RVGSharedPri 260

FileNone agentagent functions 236

attribute 236description 235resource type definition 236sample configurations 236state definitions 236

FileOnOff agentagent functions 237attribute 238description 237state definitions 238

FileOnOnly agentagent functions 239

attribute 240description 239resource type definition 240sample configuration 240state definitions 239

fire drillRVGSnaphot agent 251

Hhigh availability fire drill 30, 75, 81 , 87, 126 , 140 ,

176 , 203

Iinteger data types 18IP agent

agent functions 81attributes 82description 80high availability fire drill 81Linux attributes 82resource type definitions 85

Index



IP agent (continued)sample configurations 86state definitions 82

IPMultiNIC agentagent functions 94attributes 95description 93Linux attributes 95resource type definitions 96sample configuration 97state definitions 94

Kkeylist dimension 18

Llogownervirtual IP requirement 256

LVMLogicalVolume agentagent functions 57attributes 58description 56Linux attributes 58resource type definition 58sample configurations 58state definitions 57

LVMVolumeGroup agentagent functions 60attributes 61description 59Linux attributes 61resource type definition 61state definitions 61

Mmain.cf 18, 224main.xml 18migrating

RVGPrimary 245RVGSharedPri 260

modifyingconfiguration files 18

modifying agents 18monitor scenarios

DNS agent 126monitoring bonded NICs

Linux 91

Mount agentagent functions 65, 67attributes 68description 65high availability fire drill 75, 126 , 140Linux attributes 68notes 73resource type definition 72sample configurations 77

MultiNICA agentagent functions 102

Linux 102attributes 103description 100IP Conservation mode 101Linux

attributes 103description 100IP Conservation mode 101Performance Mode 102

Performance mode 102resource type definitions 109resource type definitions, Linux 109

NNetBIOS agent

agent functions 156

description 155resource type definition 157sample configurations 158state definitions 156

NFS agentagent functions 132

Linux 132attributes 133

Linux 133description 131resource type definition 134sample configurations 135

state definitions 133NFS lock recovery 134NFSRestart agent

agent functions 137Linux 137

attributes 139Linux 139

description 136resource type definition 139sample configuration 141

Index276



NFSRestart agent (continued)state definitions 138

NIC agentagent functions 88attributes 89

Linux 89description 86high availability fire drill 87resource type definitions

Linux 91sample configurations 92state definitions 88

noautoimport flag 31NotifierMngr agent

agent functions 214attributes 215description 213Linux attributes 215resource type definition 217sample configurations 218state definitions 214

Oonline query 126

Pparallel group

RVGShared agent 254Phantom agent

agent functions 224description 223resource type definition 224sample configurations 224

prerequisitesNFS lock recovery 134

Linux 134Samba agents 147

Process agentagent functions 204attributes 205description 203high availability fire drill 203Linux attributes 205resource type definition 206sample configurations 207state definitions 205

ProcessOnOnly agentagent functions 208

ProcessOnOnly agent (continued)attributes 208description 207Linux attributes 208resource type definition 210sample configurations 210state definitions 208

Proxy agentagent functions 220attributes 221description 220resource type definition 221sample configurations 221

R

RemoteGroup agentagent functions 226attributes 227description 225resource type definition 232state definitions 226

resource type definition 53SambaShare agent 154

resource type definitionsApplication agent 184CoordPoint agent 190DiskGroup agent 30

DiskGroupSnap agent 40DiskReservation agent 49DNS agent 123ElifNone agent 235FileNone agent 236FileOnOnly agent 240IP agent 85IPMultiNIC agent 96LVMLogicalVolume agent 58LVMVolumeGroup agent 61Mount agent 72MultiNICA agent 109

MultiNICA agent, Linux 109NetBIOS agent 157NFS agent 134NFSRestart agent 139NIC agent 91NotifierMngr agent 217Phantom agent 224Process agent 206ProcessOnOnly agent 210Proxy agent 221

Index



resource type definitions (continued)RemoteGroup agent 232SambaServer agent 152Share agent 145Volume agent 53

resourcesdescription of 18

RVG agentdescribed 241

RVGLogowner agentdependency graph 257described 256failover group 256

RVGPrimary agentdependency graph 245described 245migrating 245takeover 245

RVGShared agentdependency graph 254described 254parallel group 254

RVGSharedPri agentdependency graph 261described 260migrating 260takeover 260

RVGSnapshot agentdescribed 251fire drill 251

SSamba agents 147

overview 146prerequisites 147

SambaServer agentagent functions 149attributes 150description 149

resource type definition 152sample configuration 152state definitions 150

SambaShare agent 153agent functions 153attributes 154resource type definition 154sample configurations 154state definitions 153

sample configurationsApache HTTP server agent 171Application agent 186CoordPoint agent 192DiskGroup agent 32DiskGroupSnap agent 40DiskReservation agent 50ElifNone agent 235FileNone agent 236FileOnOff agent 238FileOnOnly agent 240IP agent 86IPMultiNIC 97LVMLogicalVolume agent 58Mount agent 77NetBIOS agent 158NFS agent 135NFSRestart agent 141NIC agent 92NotifierMngr agent 218Phantom agent 224Process agent 207ProcessOnOnly agent 210Proxy agent 221SambaServer agent 152SambaShare agent 154Share agent 146Volume agent 53

scalar dimension 18secure DNS update 127setting Mii and miimon 91Share agent 143

agent functions 143attributes 144description 142Linux attributes 144resource type definitions 145sample configurations 146state definitions 143

snapshotsusing RVGSnapshot agent for 251

State definitionsVolumeSet agent 55

state definitions 116Apache HTTP server agent 163Application agent 179CoordPoint agent 189DiskGroup agent 24DiskGroupSnap agent 34

Index278



state definitions (continued) VolumeSet agent

Index

vcs bundled agents 601 lin

Documents