vnx operating environment for file release notes · pdf file03/12/2012 · 1. emc...

1

EMC® VNX® Operating Environment

For File Version 7.1.56.5

Release Notes P/N 300-013-480

REV 03 December 03, 2012

These release notes contain information about version 7.1.56.5 of VNX OE for File. Topics include:

Revision history ......................................................................................... 2 Product description ................................................................................... 2 New features and changes ........................................................................ 2 Environment and system requirements.................................................. 4 Fixed problems ........................................................................................... 4 Known problems and limitations .......................................................... 29 Technical notes ......................................................................................... 51 Documentation ......................................................................................... 60 Software media, organization, and files ............................................... 61 Installation and upgrades ....................................................................... 62 Troubleshooting and getting help ......................................................... 67

2

Revision history

EMC VNX Operating Environment for File Version 7.1.56.5 Release Notes

Revision history

The following table presents the revision history of this document:

Revision Date Description 03 December 03, 2012 Updated for service pack release

7.1.56.5 02 September 28, 2012 Updated for service pack release

7.1.55.3 Minor editorial changes and reformatting of tables incorporated.

01 July 16, 2012 Initial release

Product description

VNX Family for unified storage

The EMC® VNX® family delivers industry-leading innovation and enterprise capabilities for File, Block, and object storage in a scalable, easy-to-use solution.

VNX series

The VNX series delivers uncompromising scalability and flexibility for the mid-tier while providing market-leading simplicity and efficiency to minimize total cost of ownership. The VNX series includes these platforms: VNX7500, VNX 5700, VNX5500, VNX5300, and VNX5100.

New features and changes

Version 7.1.56.5

VNX gateway support for new Symmetrix® Enginuity™ 5876.159.102.

• VNX gateway supports Enginuity 5876.159.102 running on any of the VMAX platforms (10K, 20K, and 40K).

• VNX gateway software features such as FAST™, TimeFinder® and SRDF®, which are supported on the earlier Symmetrix Enginuity, will be supported on 5876.159.102 Enginuity.

• VNX gateway inherits all hardware and software restrictions imposed by VMAX 10K, 20K and 40K, and Enginuity 5876.159.102.

3

New features and changes


• FAST VP with four tiers (EFD, FC, ATA and FTS) is supported on VNX gateway.

• Federated Tiered Storage (FTS) for VMAX 10K as well as for VMAX 20K and 40K, which are already supported, is also supported for the LUNs exported to VNX gateway only in case FTS devices are in FAST VP Storage groups and also the FTS is configured as an external provisioning (new FTS devices from a secondary storage, not migration or encapsulation) for the LUNs exported to VNX gateway.

• VP Compression is supported on VNX gateway. It is recommended that you use automatic compression with FAST VP or create a user defined pool for the compressed LUNs so that consistent performance can be provided for the file systems.

• I/O Quota is supported on VNX gateway. It is recommended to create a user-defined pool for the LUNs that have the same I/O Quota configuration so that the user-defined pool can provide consistent performance for file systems. If I/O Quota is set on the FAST Storage Group, there is no need to create such a user-defined pool because the FAST Storage Group is mapped to a NAS pool that can provide consistent performance for the file systems.

• The recommended configuration (best practice) for VNX gateway using I/O Quota, FAST with masking view (LUN masking feature) is as follows:

Ensure that a parent Storage Group is in a masking view with the Data Mover initiator group and Symmetrix FA port group.

Create multiple Child Storage Groups with different performance expectations. For example, extra performance, performance, and capacity.

Associate each Child Storage Group with an appropriate FAST policy.

Set an appropriate I/O Quota in each FAST Storage Group. Diskmark operation will map all FAST Storage Groups to

NAS pools. Create file systems from the mapped NAS pools.

EMC Secure Remote Support (ESRS)

This release introduces the ESRS Device Client software as part of the VNX File OE. The ESRS Device Client software provides a path for Support connect-in access to the VNX system over a secure connection, as well as traditional connect-out Callhome capability to EMC. The ESRS Device Client agent can be configured during initial Unified/File system installation, from modifications made to the VNX Installation Assistant (VIA), with the option to configure an ESRS Proxy Server connection, and ESRS Policy Manager (external to the Control Station). The VIA will also be able to configure the ESRS Device Client during the Block-to-

4

Environment and system requirements


Unified Upgrade process. Finally, upgraded/existing systems can also configure the ESRS Device Client (and optional Proxy Server, Policy Manager settings) through the VNX Unisphere® management interface. Partners/customers will still have the option to configure Email for Callhome if the ESRS Device Client agent is not going to be deployed.

Version 7.1.55.3

Blade failover/failback capability during upgrade

VNX OE for File 7.0 to 7.1 upgrade adds an option for performing failover/failback on Data Movers in addition to the currently supported option of rebooting all Data Movers.

Environment and system requirements

Symmetrix microcode and VNX for Block operating environment levels

The E-Lab™ Interoperability Navigator contains the latest microcode and VNX for Block operating environment information. The E-Lab Interoperability Navigator is updated monthly and can be found on http://Support.EMC.com. After logging in to the EMC Online Support website, locate the applicable Support by Product page, find Tools, and click E-Lab Interoperability Navigator.

Centera® FileArchiver not supported

The Centera FileArchiver policy engine is not supported when using VNX version 7.0.x.x or later.

Fixed problems

This section contains fixed problems for VNX OE for File.

Version 7.1.56.5

CIFS

Symptom description Fix summary Solution

Severity: Medium Tracking

Could not immediately connect to a network share on system when a lot of users are connecting and disconnecting

The code has been fixed to allow immediate reconnection to the share that is being

Exists in Versions: All shipping versions of 7.1 (prior to 7.1.56.5)

http://support.emc.com/

5

Fixed problems



Number: 495620 / 48114692

simultaneously. unmountedsince it is not used anymore.

Fixed in Version: 7.1.56.5

Severity: Medium Tracking Number: 499970 / 48272012

When the AV Engine was not available, the following message was logged: "Warning: RecvStream failed I/O error".

The code has been changed to suppress alert messages when killing an MSRPC client connection.

Exists in Versions: All shipping versions of 7.1 (prior to 7.1.56.5) Fixed in Version: 7.1.56.5

Severity: Medium Tracking Number: 500020

CEPP "open dir" mismatched with "file open read".

The code has been changed to use CEPP_OPEN_DIRECTORY instead of CEPP_OPEN_FILE_READ when the node is a directory.


DART Core


Severity: Critical Tracking Number: 467244 / 45564250

On file systems with a very large number of ACLs, running the ACL database printstats command took a very long time.

A “fast” option has been added to printstats command. This option provides improved response time although only a few statistics are printed.


FLR



Could not replicate FLR-enabled file systems running VNX OE for File version 7.1, with file systems running on Celerra

® NAS 6.0

versions or VNX OE for File 7.0 versions.

The code has been fixed to allow the replication of FLR-enabled file systems between 7.1 and older systems; however, enabling this requires a manual process. Please contact your


6

Fixed problems


Please see EMC Knowledgebase article emc300513 for more details.

service provider for details

Installation & Configuration

Symptom description Fix summary Solutions


Non-critical backend events were reported with CRITICAL severity in Call Home events.

The code has been fixed so that Call Home events are reported with correct severity.


NDMP



Aborting a large file system backup took a long time and the backup job appeared to be running while the system was hung.

The code has been fixed to detect and abort the backup job when the abort command is issued.


Severity: Medium Tracking Number: 480505

NDMP PAX backup could not backup and restore paths longer than 1024 bytes.

NDMP PAX backup can now backup and restore paths longer than 1024 bytes. The new limit is 4096 bytes.


NFSv2 and v3


Severity: Critical Tracking Number: 502123 /

Panic “Out of memory” with _ZN12HashLruFixedIN17NTCredentialCache21NTCred in the call traceback.

The NT credential cache is now allocated when CIFS is started if there are file systems being mounted using the ntcredential option.

Exists in Versions: All shipping versions of 7.1 (prior to 7.1.56.5) Fixed in Version:

7

Fixed problems



49412136

This avoids the panic issue of trying to allocate the cache when the system has been up long enough for memory to be fragmented.

7.1.56.5

LDAP



If a user name consisting of only one or more white space characters was used for an FTP logon, the LDAP filter processing incorrectly parsed the name, and continued parsing in memory after the end of the filter buffer. If there was a sequence of characters - starting with a backslash - in the subsequent memory, a memory corruption occurred and the Data Mover would panic.

The fix is to correctly parse the "white space" user name, to prevent the memory overrun and corruption.


NFSv4



A race condition caused the data mover to panic. The following message was displayed: DART panic/fault message: *** Page Fault Interrupt. Virt ADDRESS: 0x000062f870 Err code: 0 Target addr: 0x0000000010 **

The code has been fixed to eliminate the race condition.


8

Fixed problems




When changing UNIX mode bits with NFSv4 on a file system mounted with MIXED access policy, read_extended_attributes and write_extended_attributes might not be granted in the ACL. If ABE was enabled on the CIFS side, user might then not see files/dirs even if they had full read access to that node.

The code has been fixed to grant read_extended_attributes & write_extended_attributes correctly.



When using NFSv4, a Data Mover panic occurred due to a GP exception in nfs_wput().

The code has been fixed so that the NFSv4_setAttr could write, exceeding the mblk limit in rare conditions.


Severity: Critical Tracking Number: 503276

NFSv4 service was not responding; Pool:NFSD_Exec BLOCKED was displayed in datamover log.

The mutex-related dead-lock issue has been fixed.



NFSv4 threads were blocked.

A race condition which caused the client delegation to be registered twice, causing dead-lock when the client returned the delegation, has been fixed.


9

Fixed problems


Replication



The data mover panicked due to a memory leak condition. The following message was displayed: DART panic/fault message: >>PANIC in file: ../addrspac.cxx at line: 486 : Out of memory

The code has been fixed to eliminate the memory leak.


Security



In MIXED access mode, the user was unable to change the owner of the file although he was given WRITE_OWNER permission in ACLs.

In MIXED accesss mode, chown/chmod from a UNIX client now checks the ACL for permissions, controlled by the parameter filesystem.CheckAclInCHOps. The defualt value of 1 enables the new behavior; setting the parmeter to 0 restores the previous behavior.


SnapSure™



When replication started in reverse with HWM 0, it always returned a File_NoSpace error.

The code has been fixed so that only when the usage is greater than or equal to 75%, the replication reverse would fail, otherwise it succeeds.


10

Fixed problems



Rolling data mover panic occurred when a little amount of data wrote to PFS after a new CKPT creation and the corresponding save volume was corrupted. The following message was displayed: DART panic/fault message: >>PANIC in file: ../PagedBlockMap.cxx at line: 2423 : Duplicate Mapping

The code has been fixed to avoid the panic condition by correcting the Snapsure recovery process.


GUI


Severity: Low Frequency of Occurrence: Always Tracking Number: 497656 / 48643134

Multibyte characters were not displayed correctly (displayed as rectangles) in the File Statistics Export to File pop-up dialog.

This issue has been fixed by using the Java UIManager class to define dialog strings so they can be localized and by using the FileChooser noPlaceBar option to remove the left side menu.


Backend features


Severity: Medium Frequency of Occurrence : Always Tracking Number: 481238 /

When nas_rdf -init was done on a Control Station with host name having more than 13 ASCII characters, the following error was returned: Error 2214: <Control Station hostname> : an entry with this name already exists"

This issue has been fixed by revising the code to accept the Control Station host name to be as long as 64 ASCII characters.


11

Fixed problems


48143204

Severity: Medium Frequency of Occurrence : Always Tracking Number: 507446

NAS_DB backup failed for rdfadmin environment.

The issue has been fixed by adding environmental checks to prevent nasdmin and rdfadmin users from restoring their NAS DB from each other's DB backups.


System management


Severity: Medium Frequency of Occurrence : Rarely under a specific set of circumstances Tracking Number: 476488/ 46629630

VBB backup of VDM root file system failed.

This issue has been fixed by mounting NDMP checkpoints of VDM root file system to VDM server instead of server_2.


Severity: Low Frequency of Occurrence : Rarely under a specific set of circumstances Tracking Number: 477894/ 46797274

User would see a phantom mount point in Unisphere after using nas_replicate CLI to reverse or failover a VDM replication session.

This issue has been fixed by changing the relevant code for handling nas_replicate.

Exists in Versions: All shipping versions of 7.1 (prior to 7.1.56.5) Fixed in Versions: 7.1.56.5

Severity: Low Frequency of Occurrence :

User would see a secondary phantom mount point when enabling auto-extend on a VDM file system.

This issue has been fixed by changing relevant code for handling auto-extend


12

Fixed problems



Rarely under a specific set of circumstances Tracking Number: 504044/ 46238470

on VDM file system. Fixed in Version: 7.1.56.5

Severity: Critical Frequency of Occurrence : Sometimes Tracking Number: 486088/ 47578420 484588/ 47205760

Nas_diskmark sometimes failed to discover new LUNs with similar warning messages: Warning 17717198875: Device 0002 on APM00114900326 was skipped during diskmark operation because: device type is not set

This issue has been fixed by updating nasStorageAPI with the latest Solutions Enabler code.


Fixed in Version: 7.1.56.5

Severity: Medium Frequency of Occurrence : Always under a specific set of set of circumstances Tracking Number: 487350/ 47698828

SNMP trap did not get sent for a file system notification.

The issue was resolved by improving the code responsible for the configuration of SNMP traps.


Severity: Medium Frequency of Occurrence : Always Tracking Number:

A new mount created from GUI using all default settings got noprefetch and nonotify mount options.

The issue has been resolved by setting the correct default mount options.


13

Fixed problems



499383/ 48934908

Severity: Medium Frequency of Occurrence : Always under a specific set of set of circumstances Tracking Number: 502351/ 49425182

Failed to start fsck on a writeable checkpoint mounted on unloaded VDM.

The issue has been resolved by sending fsck request to the parent data mover of the unloaded VDM.


Severity: Low Frequency of Occurrence : Always under a specific set of set of circumstances Tracking Number: 503070 / 47564604

CIFS server created from GUI failed to join the domain due to a leading dash in administrator's password.

The issue has been resolved by adding a transient internal password prefix.


Severity: Medium Frequency of Occurrence : Always under a specific set of set of circumstances Tracking Number: 504040

Setup_clariion failed to configure 60-disk array enclosure when it was installed as primary DAE.

The issue has been resolved by adding a new set of 60-disk templates for setup_clariion.


Severity: Medium Frequency of

XML API failed to return all provisioned volumes.

The issue has been resolved by adjusting configuration parameter of Celerra


14

Fixed problems



Occurrence : Always under a specific set of set of circumstances Tracking Number: 480447/ 45516028 494164/ 47591938 495742/ 48619688

Common Cache. Fixed in Version: 7.1.56.5

Severity: Critical Frequency of Occurrence : Always under a specific set of set of circumstances Tracking Number: 495996/ 48536300

NFS export options sec=krb5p and nfsv4only were lost after modification from Unisphere GUI.

The issue has been resolved by protecting NFS exports with sec=krb5* and "nfsv4only" options from GUI modifications.


Severity: Critical Frequency of Occurrence : Always under a specific set of circumstances Tracking Number: 505417

CLI command nas_fs -info -size -all failed with error Error 2237: Execution failed: valid_item: Precondition violated. [NAS_LINKED_LIST.add])

The issue has been resolved by correcting the code error.


15

Fixed problems


VDM



Full clone across two VDMs of the same datamover could not succeed.

Support for vaai full clone across two VDMs of the same datamover added.


Version 7.1.55.3

CIFS


Severity: 2 Tracking Number: 475820 / 46376654

The user was not able to delete a DFS link that existed beneath another DFS link.

The code disallows creation of a DFS link beneath another DFS link. This is consistent with Windows server behavior.

Exists in versions: 7.1.47.5 Fixed in versions: 7.1.55.3


A file could not be created on File Level Retention (FLR) if scan on read was enabled.

The code has been updated to not scan a file if it has just been created during the open operation.



server_cifssuport -cred -path <path> command panicked the data mover when the path specified was longer than 220 characters and was unreachable.

Correctly handles unreachable path even when it is longer than 220 characters.


Severity:1 Tracking Number: 487750 / 47801326

Data Mover CPU consumption got 100% when there were a lot of file range locks.

The code has been improved to rearrange some tests in the lock manager so that checking lock conflicts consumes less CPU.


16

Fixed problems




Internet Explorer returned error when downloading files on Windows 7 with 64bit CAVA.

The code allows open for read attribute even if file is opened by another process.



OS X 10.7 could not mount CIFS FS because the reply of the cifs open may have reported invalid maximum access rights of the logged user, if rights might have been set from NFS on a file system mounted with mixed access policy.

The open reply now correctly reports maximum access rights of the logged user.



FTP to the Data Mover returned "553 - Invalid argument” error when file names with multi byte characters ending with 0x5C ( \ - DOS directory separator) were used.

The code converts SJIS filename into UTF8 before checking if it contains DOS separator in the path.



The Data Mover panicked when virus checker was enabled with param waitTimeout set. DART panic/fault message: *** Page Fault Interrupt. Virt ADDRESS: 0x000110f214 Err code: 0 Target addr: 0x0000000800 **

The code has been fixed to avoid the panic condition.


DART Core


Severity:1 Tracking Number: 471089 /

Incomplete dump on 24GB Data Mover prevented root cause analysis.

The code has been enhanced to allow settable parameter to cause panic when


17

Fixed problems



44708688

malloc heap reaches pre-defined limits.

Severity:1 Tracking Number: 483525 / 47383678

The Data Mover panicked with the following message: DART panic/fault message: >>PANIC in file: ../sched.cxx at line: 1541 : SYSTEM WATCHDOG

The code has been fixed to avoid the panic condition.


LDAP

Symptom description Fix summary Solution Severity: 2 Tracking Number: 474007 / 46106470

server_ldap reported a syntax error if bind account contained commas.

Comma in bind account is now considered as part of the account, it is no more considered as a command line separator.


NFSv4



NFSv4 access became excessively slow when there were multiple clients working on the same files.

The code has been improved to add a better locking mechanism allowing multiple requests to run in parallel, thereby improving NFSv2 access times.



NFSv4 access hung during backup.

The code fixed a deadlock related to the recall of NFSv4 file delegations.


18

Fixed problems


Quotas


Severity: 1 Tracking Number: 488023

There was no option to report capacities and usage based on file system and tree quotas that did not take user and/or group quotas into acount.

A new option has been added to the param quota useQuotasInFsStat to enable reporting only file system and tree quota accounting to df queries from Unix clients. The existing values will behave as before. 0: (default) only the file system capacities will be reported to Unix clients 1: the file system and all quotas (user, group and tree) are taken into account when capacities are reported to Unix clients 2: file system and the tree quota are taken into account when capacities are reported to Unix clients



When a large file was deleted, this was handled in the background as a deferred delete. If this process was interrupted by the file system being unmounted, it could result in tree quota accounting being incorrect.

The fix ensures the tree quota accounting is accurate.


RDF

Severity Symptom description Fix summary Solution


nas_rdf -init failed in an 8 data mover configuration when slot 9 Data Mover was selected as a standby Data Mover.

The code allows for selection of slot 9 Data Mover thereby removing the error.


19

Fixed problems



Select server(s) to use as standby: 1 2 3 4 5 6 7 8 !!!ERROR!!! server_9 has mounted user file systems. To configure it as a remote standby Data Mover, unmount all user file systems permanently (through another terminal) and try again. Select server(s) to use as standby:

Snapsure


Severity: 1 Tracking Number: 497062 / 48857254 503020 / 49495538

Repeated Data Mover reboots could be triggered during replication transfer or checkpoint merge on >2 TB file system. Any of the following messages could be displayed: Panic in file: ../DeltaIO.cxx at line: 802 : DeltaIO_Read_CB::failed Assertion failure: 'cv->unpackBtree32(curadd) >= startAddress/VOLBLK_SIZE' on line 3286 in file '../PagedBlockMap.cxx' >>PANIC in file: ../assert.c at line: 27 : Assertion failure: 'val != INVALIDNUM' on line 428 in file '../include/PagedBlo

The code has been fixed to avoid panics by:

• Avoiding the insertion of invalid block numbers

• Added checks to bypass invalid block numbers that may have been written previously.


20

Fixed problems


Severity Symptom description Fix summary Solution ckMap.hxx' >>PANIC in file: ../assert.c at line: 27 : assert failed

Storage

Symptom description Fix summary Solution Severity: 1 Tracking Number: 459091 / 44655534

A rare panic was encountered while processing a Tachyon fibre channel driver's Outbound Completion Message (OCM) with a set Invalid Entry (INV) error bit, although neither the OCM's SEST Programming Error (SPE) bit nor the VAL bit in the associated SEST were set. This condition violated the rules of the OCM's INV bit as specified in table 9-2 of the Tachyon FC Protocol Controller User Manual. The driver explicitly panicked the Data Mover (panic code = 0x01902009 == FC_EXM_INVALID_API_ERROR_STATUS with an additional error code = 0x00003002 == FC_CTLR_PROG_ERR) whenever an OCM's invalid entry bit was set without even looking at the OCM's SPE bit or the SEST entry's VAL bit.

The fix prevents this explicit panic within the Tachyon fibre channel driver (panic code = 0x01902009 == FC_EXM_INVALID_API_ERROR_STATUS) in the specific case (error code = 0x00003002 == FC_CTLR_PROG_ERR) where it violates the rules for the OCM's INV bit as specified in the Tachyon FC Protocol Controller User Manual.


21

Fixed problems


System Management



Creation of a VDM on a mirrored storage pool failed with an unclear error message: "No space left in matching storage pools".

CCMD message definition has been modified to clarify the term "matching storage pools".



Deletion of an unused disk volume from Unisphere caused underlying backend LUN to be deleted as well. The data on that LUN was lost even if the LUN was removed from the File storage group.

When deleting an unused disk volume, the code now determines if the underlying backend LUN still belongs to the File storage group. The LUN that does not belong to the File storage group will not get deleted.



File system auto extension failed when there was space in file system name.

The code quotes file system name in "rootnas_fs" command to allow spaces.



Nas_diskmark sometimes failed to discover new LUNs with a message similar to the following: Warning 17717198875: Device 0002 on APM00114900326 was skipped during diskmark operation because: device type is not set.

nasStorageAPI has been updated with the latest Solutions Enabler patch.



Jserver returned disk volumes with Symmetrix serial number that had leading zeroes.

The issue has been resolved by stripping off leading zeroes from Symmetrix serial number.


22

Fixed problems




A critical alert claiming that the Block OE software version was not compatible with the File OE software version was displayed in a system's alert list and could not be deleted. It cleared itself automatically when the cache was updated manually or at 1 a.m. by cron job so it was a temporary problem immediately following an upgrade.

A block software version check has been added to a script that runs every 5 minutes by cron job and checks for storage configuration changes and component status issues (/nas/sbin/get_backend_status). The script now invalidates the cache if the block software version changes so the alert list gets the updated value the next time the list is updated. If the versions are compatible, the critical alert will be automatically removed from a system's alert list within 10 minutes of the block software upgrade and commit.



The following error was seen in sys_log and the Unisphere event log: transfer_support_materials[26456]: The remote FTP server/directory cannot be reached. Verify your configuration in /nas/site/automaticcollection.cfg. Also, nas_checkup reported the following warning: Control Station: Check auto transfer status Warning HC_CS_18800050419: Cannot establish connection to FTP server sftphost

The issue was caused by a recent change in FTP server security, which caused the verification to fail. The code has been updated to work with the new FTP server security policy. See EMC Knowledgebase article emc296211 for the temporary workaround for fixing the problem with the script until the system is upgraded to this VNX OE for File version.


23

Fixed problems



sftp.emc.com Action : Command "/nas/tools/automaticcollection -verifyftp" failed to connect to FTP connection. Check the configuration in /nas/site/automaticcollection.cfg file. For more information, search the Knowledgebase on EMC Online Support as follows: 1. Log in to http://Support.emc.com and go to Support > Knowledgebase Search> Support Solutions Search. 2. Use ID emc221733 to search.

VDM



Interface that was attached to a CIFS server owned by a VDM still appeared as attached to the CIFS server after it was disabled.

The code has fixed the recording of disabling of such an interface in the VDM configuration file.


24

Fixed problems


Version 7.1.47.5

CIFS


Severity: 1 Tracking Number: 399822/ 39168806, 39431706, and 44143968

When the Service Principal Names (SPNs) associated with a service (for example, 'host' on a CIFS Server) were inconsistent between the Windows Active Directory and keytab in the Data Mover, client access might have been denied. This may have occurred when the server information was updated in DNS for a disjoint domain. In addition, if the user added aliases to DNS, only NTLM-type authentication would be used instead of Kerberos. This placed increased loads on the domain controller, and clients were denied access if the domain controller was not available.

For disjoint domains the previous workaround was to unjoin and rejoin the server to the domain. With this fix, the unjoin/rejoin is no longer necessary. The server_cifs command with the -setspn option allows the user to view SPNs in the Windows Active Directory and Data Mover keytab, and to update the SPNs as needed to resolve the access issue. If NTLM is used for authentication when Kerberos could have been used, the fix keeps track of these possible SPN mismatches and event logs are generated once per hour. The mismatches can also be queried using the server_cifs and server_checkup commands. This fix also allows the mismatch tracking and/or event logging to be disabled. Finally, to support troubleshooting this fix extends the server_cifssupport command so that a client IP address can be specified, for additional logon trace message in the server log.

N/A

25

Fixed problems


Severity: 1 Tracking Number: 404688/ 39080460

The usermapper exported file was not always clean if the export operation failed.

Fix has been provided to automatically remove the files if any issue occurs during the file generation.

N/A


The Data Mover hung with the following message seen in the server log: "...ThreadsServicesSupervisor: Service:LOCKD Pool:LOCKD_Exec BLOCKED for xxx seconds:..."

Fixed deadlock between lockd and statd..

N/A


Configurations with very large share databases could cause too much memory to be consumed when the Control Station enumerated the list of shares. This could result in “couldn't get a free page” panic. The probability of running into this problem increased if several concurrent enumerations were perfomed by the Control Station.

The code handling share enumeration requests from the Control Station has been redesigned to limit memory consumption.

N/A


New CIFS connections could be blocked each time the GPO Restricted Groups information was updated.

The code updating Restricted Groups is updated to not block the system while resolving the account mappings. Now Restricted Groups are first fully resolved before being applied and will not completely block LGDB access if a mapping could not be resolved.

N/A

26

Fixed problems



By design, the VNX file system did not allow the user to create more than 64k directories in a directory. On a CIFS client, when this limit was reached, the error code returned to the application was confusing (text associated to the error interpretation): "The file or directory is corrupted and unreadable" Note: this error case is not specified in Microsoft error code definition.

The code is updated so that when the limit of sub-directories you can create in a directory is reached, the error code returned to the application is more explicit: Z:\nt\homedirs>md fredt An attempt has been made to create more links on a file than the file system supports.

N/A

Deduplication



An error indicating NoSpace was logged during duplication when the primary file system had free space. The message logged looked like “…Deduplication of file system id 29 was aborted after 10 failure(s), 248409568 KB available on FS. Last reported error was: NoSpace”

The message has been clarified to indicate this is related to the deduplication FsSpaceReserve parameter setting.

N/A


The Data Mover server log was flooded with messages like “…recoverFromActiveIngest recovering (move) fs XXX”.

The severity of the log message for this common code path is lowered so it is no longer logged.

N/A

439661/ 43716310

Heavy CIFS access to a compressed file on an F-RDE-enabled filesystem could cause excessive CPU load.

Software changed to cache the file system RDE state rather than perform a more complex lookup when ever the data is needed.

N/A

27

Fixed problems


NDMP



Execution of multiple ndmpcopy operations were deadlocked and did not allow other backup operations to start.

The code is updated to avoid the deadlock.

N/A

SnapSure

Symptom description Fix summary Solution Severity: 1 Tracking Number: 395262 / 38005460, 36723806

In certain rare circumstances, checkpoint file systems might be marked as corrupt when NT event logging is enabled on the Data Mover.

The code has been updated to ensure proper locking semantics are used for writes to the windows event log.

N/A

Replication


Severity: 2 Tracking Number: 400479,410675

The Replication V2 reverse command hung and the server log contained messages like "…the command is going inactive".

A retry mechanism is implemented to start a timer to retry the replica operation later instead of letting the command become inactive.

N/A

Severity: 1 Tracking Number: 418631, 378614

After file system auto extension, destination file system size was smaller than source filesystem size, thus replication did not continue. The command, “nas_replicate -refresh” is hung.

Use block size instead of MB for file system extension, to ensure that destination file system will be extended to the same size as source file system.

N/A

Severity: 1 Tracking Number: 415849, 417590

While creating a replication or initiating a remote file system copy, some sequences of user interactions will cause a warning message to be displayed. On clicking Cancel on the warning message, the administrator selections were not retained when the control was directed back to the

The code has been updated to ensure the administrator selections will be retained when the page is repainted.

N/A

28

Fixed problems


creation page.

SRDF/S



Set mode sync operation in nas_rdf -restore failed when the Sync license was not enabled. An RDF Set 'Synchronous Mode' operation execution is in progress for device group '1R2_555_12'. Please wait... The feature being requested is not currently licensed.

The fix will ignore the license error message to continue the restore process.

N/A


nas_rdf -activate could be hung in nbs update operation. ps -ef shows "nd-clnt 6 7" like "/usr/local/bin/nd-clnt 6 7 server_2,server_3,server_4,server_5".

The fix added “sleep1” to complete the nbs stop operation before it restarts.

N/A


The nas_diskmark -mark operation removed the RDF7 & RDF10 NBS connections causing nas_rdf -init, nas_rdf -activate, and nas_rdf -restore commands to fail when masking was used. The error shown was: ERROR 13431799815 : Unable to start nbs device for /dev/ndj1

Code has been modified so that RDF7 and RDF10 nbs connections are no longer removed. The nas_rdf commands no longer fails.

N/A

UFS



A panic when deleting a file resulted in the filesystem being in an inconsistent state and lead to a panic “Unexpected state VNON in lookupComp”

The code has been updated to ensure proper transactional semantics to avoid this problem.

N/A

29

Known problems and limitations



This section includes all of the known problems found in VNX OE for File.

Celerra® Data Migration Services

System watchdog panic when using misconfigured data migration (462755)

There is a way to issue a CDMS migration in an invalid way that can also panic the Data Mover due to excessive memory consumption. Specifically, if the customer starts a CDMS file system and uses the same file system as both the source and destination it will cause this panic.

Workaround Delete the CDMS connection by issuing the halt and disconnect options.

Unisphere Service Manager

USM reports incorrect counts on number of health checks finished. (480637)

On the pre-upgrade health check step, the number of finished checks is wrong and bigger than the total number. The number of N/A, warning, and error would also be wrong.

Workaround Click Rerun to run the health check again.

Control Station does not allow Starting NAS Services while Upgrading (454458)

Powering off Control Station 0 (CS0) during a dual Control Station (Dual CS) upgrade causes failover to Control Station 1. CS0 is needed for the upgrade. Control cannot be returned to CS0 either by USM or SSH because NAS is not running on CS0 after the failover. CS1 cannot be failed over, because it requires NAS running on CS0.

Workaround Turn off NAS services on the secondary Control Station and then continue the upgrade with CLI.

Control Station

Connect Home page is disabled and ESRS doesn’t transfer events (514007)

During ESRS device client re-provision or if re-provision fails, if ESRS is configured as a transport mechanism, the Connect Home page will be disabled and ESRS won’t transfer events.

30



Workaround 1. Login to the Control Station as root.

2. Run /nas/sbin/nas_connecthome –modify –esrs_priority Disabled to disable ESRS priority.

3. Connect Home page will display the configuration and be modifiable. Switch to an alternative transport mechanism if necessary.

4. If ESRS re-provision is successful later, change back the ESRS priority.

Selecting and deleting alerts (433132)

Selecting and deleting multiple alerts does not work. Alerts must be deleted one by one.

Unisphere unresponsive (440528)

On rare occasions, the Unisphere GUI can become unresponsive if left open and idle for an extended period of time.

Workaround Open a new browser session.

Unisphere missing columns in a table (455319)

Occasionally, users might see missing columns in a GUI table if they have previously switched their language setting.

Workaround Click Reset to reset all the columns.

VIA task fails (468333)

VIA task "Primary Control Station setup" fails if the /etc/hosts secondary aliases SPA or SPB are not present.

Workaround Run "/nas/sbin/setup_clariion -check_and_update -all".

31



Unisphere no response with solid white screen (471985)

Unisphere window might show only a white screen, when left idle for long time.

Workaround 1) Restart the browser.

2) Add an environment variable JPI_PLUGIN2_NO_HEARTBEAT with any value. This prevents heartbeat monitoring and helps Unisphere stay alive.

CS Setup wizard does not support IPv6 addresses (474654)

Control Station Setup wizard does not support IPv6 addresses.

Workaround Use the Control Station Properties page instead.

Inadequate alert shows (478415)

If user uses two Unisphere clients and deletes some alerts from one client, the alert list shown in the other client does not get updated.

Workaround Restart Unisphere.

No Add/Remove buttons (480350)

The GUI will occasionally not display the Add/Remove buttons on the Add/Remove Systems dialog. This issue happens rarely and is not readily reproducible.

Workaround Restart Unisphere.

Cannot log back in after browser crash (483752)

If a browser crashes, when users try to log back in to Unisphere, the following error message is shown:

"An instance of this browser is already running Unisphere. You cannot have the application running multiple times within the same instance of a browser".

Workaround 1. Open a browser window to access "https://<application

ip>/Logout" to log out of Unisphere.

2. Log back in to Unisphere.

32



Update the AVM pool template name for R5 SAS LUNs (484620)

SAS drives are mapped to clarsas_archive and cmsas_archive system pools. The term archive is a misnomer and would apply to the slower ATA or NLSAS drives. EMC will change these pool names to clarsas_r5 and cmsas_r5 in an upcoming release.

Edit and Remove button is enabled for the Data Mover NTP server page even if there are no NTP servers configured. (485008)

If you click Edit / Remove when no NTP server is configured, an exception will be thrown, but no error message will be displayed. Additionally, the 'Add' button will become disabled.

Workaround Restart the wizard to re-enable the 'Add' button.

Replication might not properly filter targets based on the split/common log status of the source (485911)

A remote replication creation from a 7.0 common log file system to a 7.1 split log file system is not permitted and will fail with this error:

Error 13160415234: An invalid argument was specified.

Workaround Allow the system to create a destination file system from a storage pool, or ensure the source and destination file systems have the same attributes. Replications to existing file systems require that the source and destination file systems be the same size, the same file-level retention type (off, enterprise, or compliance), and the same log type (common or split).

nas_storage -info does not show director type for DX (486387)

When federated tiered storage (FTS) is configured at the backend, the output for the nas_storage -info command exhibits a missing directory type information for the DX module.

Workaround The ident column shows DX like DX-7X, if the director type is DX.

33



JServer reaching its memory limit (486639)

JServer reports reaching its memory limit, and then "out of space" periodically in /nas/jserver/logs/system_log.

Workaround Clear the odb and sdb and restart JServer.

Login to CLI as root

a) Shut down JServer: /nas/sbin/js_shutdown

b) Clear odb: /nas/sbin/js_cleanconfig

c) Clear sdb: /nas/sbin/js_cleandb

d) Start JServer: /nas/sbin/run_jserver

Battery B warning on a valid configuration on VNX5300 (487338)

The user might see an alert with warning saying one Battery (SPS) is missing. A message will be shown, such as:

Shelf 0/0 Battery B: Empty.

This hardware component is unavailable or not connected.

Check the availability of the component or the connections of the component.

Workaround This is not an error state for the system. For this type of system, there is only one power supply (battery). The alert can be ignored.

System becomes degraded when added to the domain via LDAP user (490514)

When a user logs in to Unisphere as an LDAP user and adds another system into a domain, it might take up to five minutes for the domain and LDAP configuration to propagate to the newly added system. For network reasons, it is possible that Unisphere will try to log in to the newly added system before the domain and LDAP configuration get propagated to it. This will cause login failure and result in the login session to the newly added system to be in degraded mode.

Workaround Log out of Unisphere and log in again after couple of minutes.

34



Clearing IPv6 when no IPv6 configured causes a new certificate to be generated (492605)

If IPv6 is cleared when there was no IPv6 configured, a new certificate will be generated. When a browser is opened, the user will be asked to accept the new certificate.

Workaround Accept the new certificate.

User cannot manage users after logging in as a local user. (487087)

User cannot manage users after logging in as a local user.

Workaround To manage users, log in as a global user with administrative privilege.

Changing order of DNS servers on block side does not propagate to file side (493065)

If a user changes the order of DNS servers or suffixes using Unisphere or naviseccli, the Control Station does not get updated.

Workaround The workaround is to remove all DNS settings and add them back again, or use nas_cs on the Control Station to change the order.

CLARiiON® storage system user is unauthenticated (494299)

On some integrated or Unified VNX systems, changing the backend credentials can fail with the following message:

<VNX serial number>: Storage API code=4651: SYMAPI_C_CLAR_NOT_PRIVILEGED

Operation denied by Clariion array; you cannot perform the requested operation details.

Workaround Even with the error, the backend credentials might be successfully changed. However, validation of the new credentials does not succeed and some database files are not updated. The system might report authentication errors in the GUI or when using some storage related commands for up to 5 minutes. There is a domain update that will fix the databases automatically but it might take up to 5 minutes for all files to be corrected.

35



Optional: After 5 minutes have passed, you can validate the new credentials by logging in to a fresh GUI session or SSH session with the new credentials.

Optional: After 5 minutes have passed, you can check the consistency of database files and that there are no storage related errors with the CLI command nas_storage -check -all.

Certificate is being regenerated during a gateway upgrade (495240)

During an upgrade from NFS 7.0.x.x to NFS 7.1.x.x, SSL certificates for the Control Station will be regenerated.

Workaround Accept the new certificate after the upgrade.

User cannot log in to Unisphere for several minutes after six unsuccessful login attempts (488532)

If more than six unsuccessful login attempts are made, the user's IP will be locked out of Unisphere for several minutes. Login attempts will receive a HTTP 403 error.

Workaround Wait several minutes before trying again. Check the username/password being used to ensure they are valid.

The t2reset command with the pwroff option fails to power off the Data Mover (507393)

On a control station with NAS services running, t2reset -pwroff (data mover slot) would effectively reboot the data mover (power is removed, but then restored by box monitor).

Workaround If the power needs to be removed from a data mover, you can either shut down NAS services (/sbin/service nas stop)” and then run the t2reset command, or rename /nas/sbin/recover_slot temporarily and then run the t2reset command.

Deduplication

Restoring space-reduced files

Restoring space-reduced files from a PAX-based NDMP backup into a file system causes the deduplication state to be set to suspended if the deduplication state is not currently enabled.

36



Deep compression interoperability (503537)

Files using deep compression algorithm would not be accessible on previous code families (i.e. 7.0 or lower). Therefore deep compression should not be enabled on a file system that is replicated to a Celerra/VNX running earlier code families. Refer to EMC Knowledgebase article emc303777 for more details.

ESRS

Test ESRS server connectivity before setting up ESRS Device Client (518791)

Before ESRS Device Client setup, Unisphere does not check the connectivity with EMC ESRS servers. EMC recommends that you test ESRS server connectivity before setting up the ESRS Device Client. If ESRS server connectivity is not verified, either ESRS Device Client setup from Unisphere could fail with an “unknown error” or the ESRS services may not work properly with symptoms like “Connect Home test ESRS” fails.

Workaround 1. Login to Control Station as root. 2. Run /celerra/esrs2/Configurators/check_esrs_servers to check the

EMC ESRS server connectivity. Please provide proxy server information if that is necessary to access ESRS servers.

3. If EMC ESRS servers connectivity test fails, please verify network connectivity (firewalls, router and switch across control lists and general routing).

Cannot re-provision ESRS when CS was rebooted during last ESRS provision/re-provision (521256)

When Control Station is rebooted during ESRS provision or re-provision operation, the next re-provision operation using Unisphere reports the following error:- 1. Modify ESRS Settings. Error in getting ESRS agent proxy. 2. Modify ESRS Settings. Error in getting ESRS policy manager. 3. Query ESRS Settings. Unable to connect to EMC. Please verify network connectivity or try again in a few minutes. Workaround 1. Login to Control Station as root.

37



2. Run "/sbin/service esrs2client stop" to stop esrs2client service. 3. Run "/sbin/service esrs2watchdog stop" to stop esrs2watchdog

service. 4. Re-provision ESRS on Unisphere.

Cannot configure ESRS DC through proxy with credentials (513963)

When using proxy with credentials, attempts to configure ESRC through the command line will fail with the error message “Problem accessing esrs.emc.com at port 8443”.

Workaround Use Unisphere to configure ESRS.

File-level retention

Zero sized files

Prior to Celerra version 5.6.43.9, FLR protected empty locked files from being deleted. As of version 5.6.43.9, both FLR-E and FLR-C allow the deletion of empty locked files and empty append-only files. Any locked or append-only file that contains data will be protected. This behavior will be changed in a future release such that all locked and append-only files will be protected regardless of whether they contain data.

Copying files with Windows Explorer

For systems that use version 7.0 and earlier, when using Windows Explorer to copy read-only files either into or within an FLR file system, or between FLR file systems, the protection state of the file’s copy depends on whether the Windows client making the copy uses the SMB1 or SMB2 protocol. If the Windows client uses SMB1, then the destination file will be read-only, but not protected. If the Windows client uses SMB2, the destination file will be read-only and locked with an infinite retention date. This infinite date can be reduced in an FLR-E file system, but not in an FLR-C file system.

Interoperability with Replication (501368)

Note when upgrading from 7.0 to 7.1 the FLR format of the file systems will be changed. This will make the file system incompatible with previous code families (i.e. 7.0 or lower) and prevent replication from running. If all systems involved in replication will be upgraded to 7.1 at the same time then there won’t be an issue. If there will be a delay between the systems being upgraded to 7.1 or a replication peer is not

38



able to run 7.1 then contact your service provider to discuss options for handling. Refer to EMC Knowledgebase article emc300513 for more details.

Setting up MPFS

The Set up MPFS button is available for all the systems on the Select a wizard screen. Select this option only if you want to run MPFS and the MPFS infrastructure is configured on your system in advance. Refer to the EMC VNX Series MPFS over FC and iSCSI v6.0 Linux Clients Product Guide to manually configure MPFS on your system.

Installation and upgrades

VNX for File now joined to domain during upgrade (354625)

VNX for File is joined to the VNX for Block domain automatically during an upgrade. You can use Unisphere to elevate the privileges of any user that has been transferred from the Domain to the VNX for File. This does not occur for gateway machines, however. Gateway machines cannot be a part of a VNX domain.

Control LUN requirements enforced

There is strict enforcement of the control LUN size and HLU numbers for fresh installations. An exact match between the LUN capacity and the correct HLU value is required. Mismatches will generate a "Control LUN Check Failed" error message during installation. You must correct the control LUNs before installation can complete successfully.

Error when preparing to upgrade (378606)

If the login user does not have "Control Station shell" privilege, the following error message will be displayed when selecting the File software to prepare for upgrade: Unable to connect to the provided IP address or Hostname.

Workaround Follow the Recommended Action displayed and assign "Control Station shell" privilege to the user.

Upgrade procedure overwrites /nas/site/nas_alertd.conf (506275)

Prior edits to this file are lost during upgrade. The user should be prepared to repeat any customizations that had been made prior to the upgrade.

39



Pre-upgrade health check failures during upgrade

VNX gateways cannot be upgraded while in a domain

VNX gateways running 6.0 cannot be upgraded to VNX for File 7.0 while they are in a domain. They must be removed from the 6.0 domain before being upgraded. Once installed at 7.0, the gateway cannot join another domain because 7.0 does not support the domain add/remove yet.

Multibyte support

Using multibyte characters in fields in which they are not supported.

If you use multibyte characters in data element fields that do not support multibyte characters (for example, checkpoint schedule descriptions), these fields will not display correctly. If you want these fields to be readable, you must recreate or rename the existing data element text using the ASCII character set.

Refer to EMC Knowledgebase article emc224755 for more detailed information about this limitation.

Using multibyte characters with ECC

Celerra version 5.6.40.3 and later might cause discovery and name display issues for ECC 6.x if certain VNX for File object names use non-ASCII characters. The relevant names are:

File system name

Mount point name

CIFS share name NFS export name

Quota user name

Discovery issue If non-ASCII characters are used and if the name is more than 63 characters, discovery might not work.

Garbled name display issue If these names use non-ASCII characters, the names do not display properly in ECC. However any ASCII characters used in the name do display properly. Furthermore, the name itself and the underlying data on the VNX system are not affected.

40



Suggested Workarounds Use only ASCII characters in the specified object names when using

ECC for VNX discovery. If multibyte characters must be used, keep all applicable object names under 63 characters.

Since ASCII characters display properly, append or prepend ASCII characters to object names so you can use these characters to find the correct name.

Refer to EMC Knowledgebase article emc229599 for more details about this issue.

NDMP

Incremental NDMP backups and NDMPcopy operations

Incremental NDMP backups and NDMPcopy operations use the same last backup timestamp tracking for each data set on a Data Mover. A data set in this context means the point where the backup or copy operation starts (for example, the root of a file system or a specific sub-directory within a file system). This means that if incremental NDMP backup and NDMPcopy operations of the same data set (that is, starting from the same point within a file system) are interleaved, then the following will occur:

The incremental backup will contain only changes in that data set since the NDMPcopy operation instead of since the last backup operation.

The incremental NDMPcopy of that data set will copy only those files changed since the NDMP backup instead of since the last NDMPcopy operation.

To avoid this issue, do not interleave NDMP backup operations with incremental NDMPcopy operations of the same data set; if you do, perform a full NDMP backup of that data set after any NDMPcopy activity.

Vendor support

Not all NDMP vendor versions support this version of VNX. For the latest NDMP vendors/versions/patches and restrictions supported in this version of VNX, refer to the Backup section of the E-Lab Interoperability Navigator.

41



NFS

Cached files

If a client mounts exported parent and child paths of the same file system, it is possible to access the same file from two different mount points. If client side caching is enabled, the client might end up maintaining two different versions of a cached file.

Quotas

Quota information displayed in Unisphere is sometimes incorrect (471557)

Sometimes quota information displayed in Unisphere is incorrect. Subsequent refresh of the same page returns correct information. This happens infrequently, and only when quota information is requested when the CS is rebooted.

Workaround Refreshing the quota information page corrects the problem.

Replication

Using VNX Replicator with the VDM for NFS multi-naming domain solution

The VDM for NFS multi-naming domain solution for the Data Mover in the UNIX environment is implemented by configuring an NFS server, also known as an NFS endpoint, per VDM. The VDM is used as a container for information including the file systems that are exported by the NFS endpoint. The NFS exports of the VDM are visible through a subset of the Data Mover network interfaces assigned for the VDM. The same network interface can be shared by both CIFS and NFS protocols. However, only one NFS endpoint and CIFS server is addressed through a particular logical network interface.

The exports configuration for this NFS end-point is stored in the VDM root file system. The VDM root file system is replicated as part of a VDM replication session.

Restriction The VDM for NFS multi-naming domain solution is not supported by versions of the Operating Environment earlier than 7.0.50.0. Therefore, when you replicate a VDM from a source system that runs on version 7.0.50.0 and later of the VNX Operating Environment to a destination system that runs on any version of the Operating Environment earlier than 7.0.50.0, the file systems replicate correctly, but the NFS endpoint does not work on the system with the Operating Environment earlier than version 7.0.50.0.

42



Replication Session exists even after RP failover (467550)

The nas_copy command might not complete after rdf failover. On the file cabinet DR remote site, when logged in as any administrator other than DRAdmin, from Unisphere and NAS CLIs, the user can see that the copy operation is still in-progress but making no progress.

Workaround There are two options to work around this problem.

As a non- DRAdmin user on the remote VNX, the user can choose to locally abort the copy operation.

The user can issue DR failback and the copy should finish by itself after the DR failback.

Security

Changing LDAP settings (374023)

When running the server_netstat command, or updating the LDAP settings, the following error was displayed:

The lockbox stable value threshold was not met because the system fingerprint has changed. Reset the SSV values by running "/nas/sbin/cst_setup -reset" from the CLI to resolve the issue.

SnapSure

Parameter file

Do not insert a new-line character at the end of a nas_param file (/nas/site/ or /nas/sys/). The empty line causes fs_ckpt to return an encodeable.extract error.

Writable checkpoints operations

Writable checkpoints do not support the following operations and features:

Checkpoint refresh

Checkpoint schedule

VNX replication

CDMS File-level retention

File system extension

iSCSI

43



MirrorView™/SRDF

TimeFinder™/RDF

VTLU

SRDF/S

nas_rdf –activate

The nas_rdf –activate –reverse command might fail to swap device-pair personalities during the activate. This occurs when the source VNX has a TimeFinder/FS snapshot/s created on R1BCVs that use TimeFinder Emulation mode (Symm 8 supports Emulation mode only). The error message below is specific to the personality swap and having remotely protected storage, and does not impact the successful activation of the destination VNX system.

Error message in nas_rdf –activate –reverse:

An RDF 'Swap Personality' operation execution is in progress for device group '1R2_500_3'. Please wait...

Cannot use the device for this function because it is a Copy Session target

ERROR 13431799810 : Swap operation of device group 1R2_500_3 failed due to the improper backend state or configuration.

Run 'nas_message -info 13431799810' to get full description and recommended action.

Workaround Canceling the relationship between the R1BCV mirrored volumes and the R1STD volumes is needed to complete the swap operation.

After the SRDF restore completes, the SYMAPI will automatically request a full establish during the subsequent fs_timefinder -M refresh command to the snapshots built on the canceled R1BCV volumes.

Table 1 nas_rdf –activate –reverse procedure Step Action 1 To check symm id of source Symmetrix, run /nas/symcli/bin/symcfg list on the

destination SRDF system. 2 To check the mirrored volumes, run /nas/symcli/bin/symmir –sid <source symm id> list

on the destination SRDF system. You can also double check if Emulation mode is used from the output.

3 To find R2 device group, run /nas/symcli/bin/symdg list on the destination SRDF system.

4 To find R1BCV volumes, run /nas/symcli/bin/symrdf –g <r2 group name> query –bcv on the destination SRDF system. If the same device ID is seen in the output, that is the R1BCV device you need to cancel.

44



5 To prepare for the cancel operation, create a file which has all R1BCV volume pairs to cancel on destination SRDF system. Example of the file: # cat file # <standard device id><bcv device id> 0068 00B8 0069 00B9

6 To cancel the R1BCV volumes, run /nas/symcli/bin/symmir –sid <source sid> -f <file name> cancel on the destination SRDF system.

7 To confirm the canceled volume, run /nas/symcli/bin/symmir –sid <source symm id> list on the destination SRDF system. The canceled volumes should not be listed in the output.

8 To compete the activate reverse operation, run /nas/sbin/nas_rdf –activate –reverse on the destination SRDF system.

nas_rdf –restore fails

The nas_rdf –restore command might fail on the SRDF system, which uses TimeFinder Emulation mode (Symm 8 supports Emulation mode only) with R2BCV mirrored volumes.

Error message in nas_rdf –restore:

An RDF 'Failover' operation execution is n progress for device group '1R2_500_3'. Please wait...

Cannot use the device for this function because it is a Copy Session target

CRITICAL FAULT: SRDF restore failed for 1R2_500_3

Workaround Canceling the relationship between the R2BCV mirrored volumes and the R2STD volumes is needed to complete the restore operation. After SRDF restore completes, full establish is required when we turn on mirroring again by using fs_timefinder to the canceled R2BCV volumes.

Table 2 nas_rdf –activate –restore procedure Step Action 1 To check the mirrored volumes, run /nas/symcli/bin/symmir –sid list on the destination

SRDF system. You can also double check if Emulation mode is used from the output.

2 To find R2 device group, run /nas/symcli/bin/symdg list on the destination SRDF system.

3 To find R2BCV volumes, run /nas/symcli/bin/symrdf –g <r2 group name> query –bcv on the destination SRDF system. If the same device ID is seen in the output, that is the R2BCV device you need to cancel.

4 To cancel the R2BCV volumes, run /nas/symcli/bin/symmir –g <r2 group name> cancel -force on the destination SRDF system.

45



Step Action 5 To confirm the canceled volume, run /nas/symcli/bin/symmir –sid list on the destination

SRDF system. The canceled volumes should not be listed in the output.

6 To compete restore operation, run /nas/sbin/nas_rdf –restore on the destination SRDF system.

Nas_rdf –init command fails

During the SRDF discover remote storage devices process, the nas_rdf -init command fails and returns the following error: "Unable to setup remote devices".

Workaround Delete the device group for R2 devices using /nas/symcli/bin/symdg delete <DgName> -force.

For example: /nas/symcli/bin/symdg delete 1R2_500_13 –force. Then run nas_rdf –init, to re-create the R2 device group.

EFD drives not supported

SRDF/S does not support EFD drives. If you attempt to configure an EFD drive with SRDF/S, VNX health check operations reject the device as an unsupported storage. However SRDF or Timefinder support R1EFD, R2EFD and BCV EFD thorough FAST™ Storage Group. These devices must be added to the FAST Storage Group to be used on VNX for file systems. nas_rdf –init fails with Error 2214: an entry with this name already exists (491238)

The following error is seen when nas_rdf -init is done on the Control Station whose host name is more than 13 ASCII characters:

Error 2214: <Control Station hostname> : an entry with this name already exists. Workaround

Make the Control Station host name shorter than 14 ASCII characters.

SRDF STAR Timefinder diskmark restore issue

nas_diskmark will fail in the step to restore diskmarks to BCVs on Site B or Site C.

# nas_diskmark -m -upgrade -a

Enter password: 4fd8da85

Error 5002:

46



conflict:storageID-devID: disk=77 old:symm=000194900462,dev=10A4 new:symm=000194900462,dev=10A4 addr=c15t9l0

conflict:storageID-devID: disk=78 old:symm=000194900462,dev=10A5 new:symm=000194900462,dev=10A5 addr=c15t9l1

Workaround

A BCV device on Site B or Site C needs to use a different host ID from the one used on a R1 device pair on Site A to avoid this issue.

For example, if R1 devices are using 0x10 - 0x20, BCV devices should use 0x21 - 0x 31 etc.

Unisphere

The following notes apply to Unisphere.

No storage system value displayed (460767)

When all columns are displayed in the Storage Pool page, no values are shown for the storage system.

Workaround Storage system values can be displayed through the use of the CLI.

Browsers

EMC requires the use of one of the following browsers to run Unisphere:

Internet Explorer 6.0 SP2, 6.2.3 through 8.0; Internet Explorer 7.1 is recommended.

Firefox 3.0 or later.

Firefox Firefox Password Manager (PM) is enabled by default. This feature records and stores the password that you enter on a website and then automatically fills in the password entry field on your next and subsequent visit. The password field masks the password with asterisks (*).

To prevent Firefox from remembering passwords:

1. Click Tools > Options... and select the Security tab.

2. Under Passwords, clear the Remember password for sites check box.

47



List page not refreshed (396845)

Occasionally, Unisphere does not refresh the list page automatically after creating or deleting operation.

Workaround Refresh the page using the green arrow icon.

Popup pages go to background (396885)

Unisphere pop-up pages go to the background and the user needs to bring the pop-up window to the foreground as the active window to make it visible.

Workaround This is a browser compatibility issue. On Windows, click on the new browser window representation in the Windows task bar, use Alt-tab to step through the active windows and choose the new browser window, or use Windows Task Manager to make the new browser window active. You can also try other browsers, such as IE 8.0.6001, which does not have the problem.

Changes from other user interfaces

Any changes made to the network server using another user interface such as the CLI are not automatically shown in Unisphere. Refresh the browser to see these changes.

Confirmation dialog inactive (37974)

While using Unisphere, a confirmation dialog appears briefly, then moves to the background. Users need to manually re-focus the window and bring the dialog box to the foreground.

Dashboard summary (364539) The Dashboard Summary page does not contain view block. This is unexpected, but those blocks can be re-added using the Customize button.

File system delete confirmation (372252)

The OK button is disabled on the Unisphere File System Delete Confirmation page for a file system with file-level retention enabled at the compliance level. The table on the File System Delete Confirmation page indicates that protected files still exist on the file system even though the File System Properties page and the 'nas_fs -i' CLI command both show that all file protection has expired for the file system. The user is unable to delete the file system using Unisphere.

48



Workaround The File-level retention status is now cached for performance reasons. Force the file retention data to be updated by selecting the Unisphere Refresh for the File System List page (in the upper right corner of the File System List panel) or by opening a new browser session.

Logging in via multiple screens (163075)

Logging into VNX through two or more screens might succeed, but logging out of one of the sessions causes the other sessions to lose contact.

Login screen (370340)

When you log into Unisphere through a VNX system, Unisphere will not display a login screen if you click yes on the security warning dialog, without also checking the Always trust content checkbox on the same dialog.

Storage domain (376823)

When the VNX belongs to a storage domain that is destroyed, the storage domain users on the VNX Control Station are not automatically moved to the "defunct" state. To correct this, manually remove all the storage domain users from the system by running the command:

/nas/http/webui/bin/admin_management.pl - defunctStorageUsers

Then remove the VNX from the domain.

User’s tab displays unrelated messages (364513)

Unisphere > Settings > User's tab displays messages that are unrelated to that screen. These messages are a result of a query that was initiated by the user on some previous table screen.

Workaround Visiting another table screen and then to coming back to the table screen where the messages were seen should remove these messages.

VNX Installation Assistant for File/Unified (VIA, formerly CSA)

VIA might time out when configuring primary CS with version 7.0.51.3 (471557)

This issue might occur after a fresh install. On the Applying page, when executing the task:

Configuring Primary Control Station,

49



VIA reports the error:

Re-authentication:

Connection to VNX has timed out.

Username: root

Password:

Workaround When this error occurs the user should enter the root password at the prompt to continue the task.

VNX for Block storage pools not displayed (397788)

VNX for Block storage pools mapped to the file side are not displayed in the VIA CIFS wizard.

Workaround Use Unisphere instead to create file systems from these mapped pools.

Zero time remaining during disk provisioning process (396228)

VIA incorrectly displays zero time remaining during the LUN marking task part of the disk provisioning process. If this occurs, continue to wait for this task to complete because the time remaining estimate is incorrect.

Caution

Do not install Linux non-VNX applications on the Control Station. Doing so consumes resources and might adversely affect system operation. VNX Control Station is a dedicated controller and not an applications platform.

Do not manually edit the nas_db database without consulting EMC Customer Service. Any changes to this file could cause problems with your VNX installation.

Command line scripts

The NAS database stores specific information required for each Data Mover. The VNX for file automatically performs a backup of the entire database every hour and saves it to a file named nasdb_backup.1.tar.gz. These backups start one minute after the hour.

EMC recommends that you check your scripts to make sure they are not scheduled to start at the same time as the backups. During backup, the database becomes locked or inaccessible, and some commands that rely on the database might fail.

50



VIA fails to successfully change the IP addresses to public IP addresses after the sysadmin password has been changed from its default. (495434)

If an installer changes the sysadmin array account (optional but encouraged for security reasons) , then VIA might fail when attempting to change the storage system IP address (not optional).

The VIA task "Storage system IP address and System Name configuration" task fails with the following detailed error message:

Severity: Error

Message: The sysadmin account for the storage system does not exist.

Full Description: The backend IP address and name could not be set up, or the Proxy ARP check failed.

Recommended action: Verify if the sysadmin account exists on the storage system. If the sysadmin account does not exist, create it and retry the operation.

Workaround The user should wait approximately 20 minutes for the new password to propagate and not attempt accessing the VNX backend array with an invalid credential. The array and CS will clear after this time and the operation can be restarted.

server_http

The server_http command fails if you issue it on a Data Mover while running fsck. Other commands (including FileMover API) might not work until fsck completes.

Virtual device naming

When creating virtual devices, such as Fail-Safe Network (FSN) devices, Ethernet channels, and link aggregations, do not use the following names as the name of the device:

Amgr http mac pfkey spstream Tcprpc

Blksrv i82596 Meshforwarder Pipe Ssl Tftp Client Icmp Nd Pmap Sslpipe Toeframechk Cmfap icmp6 Nfs Rcp Statd Upd Dep Ip nfsPrimary Rfa StreamHead Udpnfsmesh Dfe ip6 Nwh Scsicfg Tcp

Echo Lockd Pax Smb Tcpnfsmesh Escon Lpfn Pcnfs Smpte tcprpc

51

Technical notes


Technical notes

This section includes the technical information for VNX OE for File.

Configuration guidelines

The values listed in this section represent the highest or lowest values tested by EMC. In some cases, actual capacities might exceed the capacities tested by EMC. Theoretical means in theory, not always practical. Refer to the E-Lab Interoperability Navigator for the most up-to-date information regarding these guidelines.

Table 3. CIFS guidelines Guideline/ Specification

Maximum tested value

Comment

CIFS TCP connection

64K (default and theoretical max.), 20K (max. tested)

Param tcp.maxStreams sets the maximum number of TCP connections a Data Mover can have. Maximum value is 64K, or 65535. TCP connections (streams) are shared by other components and should be changed in monitored small increments.

Share name length

80 characters (Unicode). 12 chars (ASCII)

Unicode: The maximum length for a share name with Unicode enabled is 80 characters. ASCII: In ASCII mode, the maximum share length is 12 characters.

Number of CIFS shares

10,000 per Data Mover 40,000 per VNX for File

You might notice a performance impact when managing a large number of shares.

Number of NetBIOS names/compnames per Data Mover

509 (max) Limited by the number of network interfaces available on the Data Mover. From a local group perspective, the number is limited to 509. NetBIOS and compnames must be associated with at least one unique network interface.

NetBIOS name length

15 NetBIOS names are limited to 15 characters (Microsoft limit) and cannot begin with an@ (at sign) or a - (dash) character. The name also cannot include white space, tab characters, or the following symbols: / \ : ; , = * + | [ ] ? < > “. If using compnames, the NetBIOS form of the name is assigned automatically and is derived from the first 15 characters of the <comp_name>.

Comment (ASCII chars) for NetBIOS name for server

256 Limited to 256 ASCII characters. • Restricted Characters: You cannot use double quotation ("), semicolon (;), accent (`), and comma (,) characters within the body of a comment. Attempting to use these special characters results in an error message. You can only use an exclamation point (!) if it is preceded by a single quotation mark (’). • Default Comments: If you do not explicitly add a comment, the system adds a default comment of the form EMC-

52

Technical notes


Guideline/ Specification


Comment

SNAS:T<x.x.x.x> where <x.x.x.x> is the version of the NAS software.

Compname length

63 bytes For integration with Windows environment releases later than Windows 2000, the CIFS server computer name length can be up to 21 characters when UTF-8 (3 bytes char) is used.

Number of domains

10 tested 509 (theoretical max.) The maximum number of Windows domains a Data Mover can be a member of. To increase the max default value from 32, change parameter cifs.lsarpc.maxDomain. The Parameters Guide for VNX for File contains more detailed information about this parameter.

Block size negotiated

64 KB 128KB with SMB2

Maximum buffer size that can be negotiated with Microsoft Windows clients. To increase the default value, change param cifs.W95BufSz, cifs.NTBufSz or cifs.W2KBufSz. The Parameters Guide for VNX for File contains more detailed information about these parameters. Note: With SMB2.1, read and write operation supports 1MB buffer (this feature is named “large MTU”).

Number of simultaneous requests per CIFS session (maxMpxCount)

127 This value is fixed and defines the number of requests a client is able to send to the Data Mover at the same time (for example, a change notification request). To increase this value, change the maxMpxCount parameter. The Parameters Guide for VNX for File contains more detailed information about this parameter.

Total number files/directories opened per Data Mover

200,000 A large number of open files could require high memory usage on the Data Mover and potentially lead to out-of-memory issues.

Number of Home Directories supported

20,000

Number of Windows/UNIX users

Earlier versions of the VNX server relied on a basic database, nameDB, to maintain Usermapper and secmap mapping information. DBMS now replaces the basic database. This solves the inode consumption issue and provides better consistency and recoverability with the support of database transactions. It also provides better atomicity, isolation, and durability in database management.

Number of users per TCP connection

64K To decrease the default value, change param cifs.listBlocks (default 255, max 255). The value of this parameter times 256 = max number of users. Note: TID/FID/UID shares this parameter and cannot be changed individually for each ID. Use caution when increasing this value as it could lead to an out-of-memory condition. Refer to the Parameters Guide for VNX for File for parameter

53

Technical notes




Comment

information.

Number of files/directories opened per CIFS connection in SMB1

64K To decrease the default value, change param cifs.listBlocks (default 255, max 255). The value of this parameter times 256 = max number of files/directories opened per CIFS connection. Note: TID/FID/UID shares this parameter and cannot be changed individually for each ID. Use caution when increasing this value as it could lead to an out-of-memory condition. Be sure to follow the recommendation for total number of files/directories opened per Data Mover. Refer to the System Parameters Guide for VNX for File for parameter information.

Number of files/directories opened per CIFS connection in SMB2

127K To decrease the default value, change parameter cifs.smb2.listBlocks (default 511, max 511). The value of this parameter times 256 = max number of files/directories opened per CIFS connection.

Number of VDMs per Data Mover

29 The total number of VDMs, file systems, and checkpoints across a whole cabinet cannot exceed 2048.

number of ntxmap User Mapping entries

1000

Table 4. FileMover



comment

Max connections to secondary storage per primary (VNX for File) file system

1024

Number of HTTP threads for servicing FileMover API requests per Data Mover

64 Number of threads available for recalling data from secondary storage is half the number of whichever is the lower of CIFS or NFS threads. 16 (default), can be increased using server_http command, max (tested) is 64.

Table 5. File systems guidelines



comment

Mount point name length

255 bytes (ASCII) The "/" is used when creating the mount point and is equal to one character. If exceeded, Error 4105: Server_x:path_name: invalid path specified is returned.

File system 240 bytes (ASCII) For nas_fs list, the name of a file system will be

54

Technical notes




comment

name length 19 chars display for list option

truncated if it is more than 19 characters. To display the full file system name, use the –info option with a file system ID (nas_fs -i id=<fsid>).

Filename length

255 bytes (NFS), 255 characters (CIFS)

With Unicode enabled in an NFS environment, the number of characters that can be stored depends on the client encoding type such as latin-1. For example: With Unicode enabled, a Japanese UTF-8 character may require three bytes. With Unicode enabled in a CIFS environment, the maximum number of characters is 255. For filenames shared between NFS and CIFS, CIFS allows 255 characters. NFS truncates these names when they are more than 255 bytes in UTF-8, and manages the file successfully.

Pathname length

1,024 bytes Note: Make sure the final path length of restored files is less than 1024 bytes. For example, if a file is backed up which originally had path name of 900 bytes, and it is restoring to a path with 400 bytes, the final path length would be 1300 bytes and would not be restored. 1

Directory name length

255 bytes This is a hard limit and is rejected on creation if over the 255 limit.The limit is bytes for UNIX names, Unicode characters for CIFS.

Subdirectories (per parent directory)

65,533 This is a hard limit, code will prevent you from creating more than 65,533 directories.

Number of file systems per VNX

4096 This max number includes VDM and checkpoint file systems.

Number of file systems per Data Mover

2048 The mount operation will fail when the number of file systems reaches 2048 with an error indicating maximum number of file systems reached. This maximum number includes VDM and checkpoint file systems.

Maximum disk volume s ize

Dependent on RAID group (see comments)

Unified platforms — Running setup_clariion on the VNX7500 platform will provision the storage to use 1 LUN per RAID group which might result in LUNS > 2 TB, depending on drive capacity and number of drives in the RAID group. On all VNX integrated platforms (VNX5300, VNX5500, VNX5700, VNX7500), the 2 TB LUN limitation has been lifted--this might or might not result in LUN size larger than 2 TB depending on RAID group size and drive capacity. For all other unified platforms, setup_clariion will continue to function as in the past, breaking large RAID groups into LUNS that are < 2 TB in size.

55

Technical notes




comment

Gateway systems — For gateway systems, users might configure LUNs greater than 2 TB up to 16 TB or max size of the RAID group, whichever is less. This is supported for VG2 and VG8 NAS gateways when attached to CX3, CX4, VNX, and Symmetrix DMX™-4 and VMAX

® backends.

Multi-Path File Systems (MPFS)— MPFS supports LUNs greater than 2 TB. Windows 2000 and 32-bit Windows XP, however, cannot support large LUNs due to a Windows OS limitation. All other MPFS Windows clients support LUN sizes greater than 2 TB if the 5.0.90.900 patch is applied. Use of these 32-bit Windows XP clients on VNX7500 systems require request for price quotation (RPQ) approval.

Total storage for a Data Mover (Fibre Channel Only)

VNX Version 7.0 200 TB VNX5300 256 TB VNX5500, VNX5700, VNX7500, VG2 & VG8 Celerra Version 6.0 16 TB = NX4 32 TB =NS20, NS- 120 64 TB =NS40, NS- 480, NS40G, NS-G2 128 TB = NS80, NS80G, NSX 256 TB = NS960, NS-G8 Celerra Version 5.6 8 TB = CNS (514 DM) 10 TB = NX350, NS500 16 TB = NX4, NS700 32 TB = NS20, NS120 64 TB = NS40, NS480, NS40G, NS-G2, VG2 128 TB = NS80, NS960, NS80G, NSX, NS-G8

These total capacity values represent Fibre Channel disk maximum with no ATA drives. Fibre Channel capacity will change if ATA is used. Notes:

• On a per-Data-Mover basis, the total size of all file systems, and the size of all SavVols used by SnapSure must be less than the total supported capacity. Exceeding these limit can cause out of memory panic.

• Refer to the VNX for file capacity limits tables for more information, including mixed disk type configurations.

56

Technical notes




comment

File system size

See right column for TB information per Data Mover

16 TB (hard limit, enforced) NX4, NS20, NS120, NS40, NS480, NS80, NS960, NS40G, NS-G8, NS80G, NSX, NS-G8, VG2, VG8, and all of VNX series 10 TB NS350 8 TB 514/514T Data Mover, NS500 When planning a file system, with respect to maximum desired size, be sure to factor in the time required for fsck, ACL checks, file system build and restore. This guideline applies to both Fibre Channel and ATA drives.

File size 16 TB This hard limit is enforced and cannot be exceeded. Number of directories supported per file system

Same as number of inodes in file system. Each 8 KB of space = 1 inode.

Number of files per directory

500,000 Exceeding this number will cause performance problems.

Maximum number of files and directories per VNX file system

256 million (default) This is the total number of files and directories that can be in a single VNX file system. This number can be increased to 4 billion at file system creation time but should only be done after considering recovery and restore time and the total storage utilized per file. The actual maximum number in a given file system is dependent on a number of factors including the size of the file system.

Maximum amount of deduplicated data supported

256 TB When quotas are enabled with file size policy, the maximum amount of deduplicated data supported is 256 TB. This amount includes other files owned by UID 0 or GID.

All other industry-standard caveats, restrictions, policies, and best practices prevail. This includes, but is not limited to, fsck times (now made faster through multi-threading), backup and restore times, number of objects per file system, snapshots, file system replication, VDM replication, performance, availability, extend times, and layout policies. Proper planning and preparation should occur prior to implementing these guidelines.

Table 6. Naming services guidelines Guideline/Specification Maximum tested

value comment

Number of DNS domains 3 - WebUI unlimited – CLI

Three DNS servers per Data Mover is the limit if using WebUI. There is no limit when using the command line interface (CLI).

Number of NIS Servers 10 You can configure up to 10 NIS servers in a

57

Technical notes


Guideline/Specification Maximum tested value

comment

per Data Mover single NIS domain on a Data Mover.

NIS record capacity 1004 bytes A Data Mover can read 1004 bytes of data from a NIS record.

Number of DNS servers per DNS domain

3

Table 7. NFS guidelines

Guideline/ Specification Maximum tested value

comment

Number of NFS exports 2,048 per Data Mover tested Unlimited theoretical max.

You might notice a performance impact when managing a large number of exports using Unisphere.

Number of concurrent NFS clients

64K with TCP (theoretical) Unlimited with UDP (theoretical)

Limited by TCP connections.

Netgroup line size 16383 The maximum line length that the Data Mover will accept in the local netgroup file on the Data Mover or the netgroup map in the NIS domain that the Data Mover is bound to.

Number of UNIX groups supported

64K 2 billion max value of any GID. The maximum number of GIDs is 64K, but an individual GID can have an ID in the range of 0- 2147483648.

Table 8. Networking guidelines



comment

Link aggregation/ether channel

8 ports (ether channel) 12 ports (link aggregation (LACP)

Ether channel: the number of ports used must be a power of 2 (2, 4, or 8). Link aggregation: any number of ports can be used. All ports must be the same speed. Mixing different NIC types (that is copper and fibre) is not recommended.

Number of VLANs supported

4094 IEEE standard.

Number of interfaces per Data Mover

45 tested Theoretically 509.

Number of FTP connections

Theoretical value 64K

By default the value is (in theory) 0xFFFF, but it is also limited by the number of TCP streams that can be opened. To increase the default value, change param tcp.maxStreams (set to 0x00000800 by default). If you increase it to 64K before you start TCP, you will not be

58

Technical notes




comment

able to increase the number of FTP connections. Refer to the Parameters Guide for VNX for File for parameter information.

59

Technical notes


Table 9. Quotas guidelines Guideline/Specification Maximum tested value Comment Number of tree quotas 8191 Per file system

Max size of tree quotas 256 TB Includes file size and quota tree size.

Max number of unique groups 64K Per file system

Quota path length 1024 Table 10. Replicator V2 guidelines

Guideline/Specification Maximum tested value Comment Number of replication sessions per VNX

1365 (NSX) 682 (other platforms)

Max number of r eplication sessions per Data Mover

1024 This enforced limit includes all configured file system, VDM, and copy sessions.

Max number of local and remote file system and VDM replication sessions per Data Mover

682

Max number of loopback file system and VDM replication sessions per Data Mover

341

Table 11. Snapsure guidelines

Guideline/Specification Maximum tested value Comment Number of checkpoints per file system

96 read-only 16 writeable

Up to 96 read-only checkpoints per file system are supported as well as 16 writeable checkpoints.

On systems with Unicode enabled, a character might require between 1 and 3 bytes, depending on encoding type or character used. For example, a Japanese character typically uses 3 bytes in UTF-8. ASCII characters require 1 byte.

Capacity guidelines

VNX for File capacity limits

Following is a list of the version 7.0 VNX for file capacity limits. Refer to the E-Lab™ Interoperability Navigator, which is updated monthly, for the most up-to-date capacity limits.

60

Documentation


Table 12 VNX for File version 7.0 capacity limit VNX5300 VNX5500 VNX5700 VNX7500 VG2 VG8

Usable IP storage capacity limit per blade (all disk types and uses) 3

200 TB 256 TB 256 TB 256 TB 256 TB 256 TB

Max FS size 16TB 16TB 16TB 16TB 16TB 16TB Max # FS per DM/Blade 4

2048 2048 2048 2048 2048 2048

Max # FS per cabinet 4096 4096 4096 4096 4096 4096 Max configured replication sessions per DM/blade for Replicator v2 2

1024 1024 1024 1024 1024 1024

Max # of checkpoints per PFS 1

96 96 96 96 96 96

Max # of NDMP sessions per DM/blade

4 8 8 8 4 8

Memory per DM/blade 6GB 12GB 12GB 24GB 6GB 24GB 1. PFS is production file system 2. A maximum of 256 concurrently transferring sessions are supported per DM 3. This is the usable IP storage capacity per blade. For overall platform capacity, consider the type and size of

disk drive and the usable Fibre Channel host capacity requirements, RAID group options and the total number of disks supported

4. This count includes production file systems, user-defined file system checkpoints, and two checkpoints for each replication session.

Documentation

VNX OE for File version 7.1 documentation

EMC provides the ability to create step-by-step planning, installation, and maintenance instructions tailored to your environment. To create VNX customized documentation, go to: https://mydocs.emc.com/VNX.

Documentation updates

Multibyte support for VNX data elements

In the VNX management interfaces, multibyte support allows users to input, store, and display multibyte characters. Users can input and view VNX object fields in their native language. The multibyte feature is supported since Celerra NAS 5.6 and later.

https://mydocs.emc.com/VNX

61

Software media, organization, and files


The following data elements are enabled with multibyte support and also support the Unicode 3.0 standard:

CIFS share name

CIFS share comment CIFS export path

Tree quota path

Tree quota comment

Quota user name

CIFS server netbios name

CIFS server computer name CIFS server alias

CIFS server comment

CIFS domain name

CIFS workgroup

Organizational unit (OU) name

DFS root name CDMS source share name

CDMS local path

CDMS source export path

Mount point name

File system name

Checkpoint name NFS export name

NFS export path

Software media, organization, and files

The following table lists the applicable software media for this product version:

Table 13. Available software Part number Description 453-005-676 VNX FILE SOFTWARE 7.1.X DVD IMAGE

62




VNX for File installation process

This section lists some facts you should know about the VNX for File and the installation and configuration process before proceeding further.

How to prepare for VNX for File installation

To prepare for VNX for File installation and configuration, complete the following tasks:

1. Perform the pre-site installation tasks as required by EMC to complete the installation and connectivity of the VNX for file. Tasks include preparing electrical facilities, wiring the network, adding a telephone line, checking environmental conditions and related requirements, as stated in the Professional Services Statement of Work (PSSOW).

Note: For a copy of the PS-SOW document, contact your Professional Services Account Team. For examples of VNX for File network topologies, refer to the Configuring and Managing Networking for VNX technical module.

2. Configure your environment for UNIX clients, Microsoft Windows clients, or a combination of Windows and UNIX clients, as applicable.

3. Complete network assessment to ensure proper performance and protocol configurations. (Optional EMC services can be ordered to assist with this process.)

VNX for File installation and configuration

EMC Customer Service or the customer installs the VNX for File hardware a process that includes the physical setup, cabling, power application and diagnostic testing of the VNX for File.

Next, EMC Professional Services or an EMC partner might configure and integrate the VNX for file. For example, for a VNX for File with a Symmetrix system, this process typically includes the following tasks:

Coordinating the activities to enable installation, cabling, and connectivity to system cabinets

Loading and configuring the VNX for File software

Creating, mounting, and exporting file systems

Configuring network protocols

63



Integrating with a supported, existing user authentication environment

Configuring standby Data Mover(s), if applicable

Configuring and demonstrating system management and administration tools

EMC Professional Services can perform additional VNX for File configuration tasks, such as host provisioning (with VNX for File MPFS), backup protection, and network connectivity, depending on the individual EMC customer agreement defined in the PSSOW. These services are part of the EMC Network-Attached Storage Design and Implementation service offering, which is designed to provide continuous technical expertise as you manage your organization’s growing network storage needs. For more information about this service offering, contact your EMC Customer Support representative.

After EMC installs, configures, and tests the VNX for File, it is fully operational and ready to support networked storage operations.

NAS and VNX for File and Block operating environment upgrades

EMC requires VNX for File upgrades to be performed before the VNX for Block operating environment is upgraded on any attached VNX for Block systems.

The reason for this is to maintain compatibility between the storage API on the Control Station and the VNX for Block operating environment. The API is part of the VNX for File management software and is used to gather data about the backend such as statistics, health, and configuration.

If the VNX for Block operating environment is upgraded first, the API might encounter issues while communicating with the new VNX for Block operating environment. Upgrading VNX for File first allows the API to have limited backward compatibility with the VNX for Block operating environment code.

Reboot required at upgrade

VNX OE for File The VNX OE for File software is comprised of two installation components, a Linux package and a VNX OE for File software package. The software currently running on the VNX depends on the File upgrade software versions to determine the number of Control Station and Data Mover reboots required and whether Data Mover reboots can be scheduled to a later time after the upgrade is complete.

The VNX for File software runs on the Data Movers and is part of the VNX OE for File software package. The new version of VNX for File

64



operating system is loaded and run on the Data Movers only after a Data Mover boot.

Control Station rebooting Most software upgrades require a Control Station reboot. When upgrading VNX for File software, if either of the first two digits of the new version number change (for example, 6.0.x to 7.0.x), the Control Station requires rebooting.

At the beginning of upgrade, the system information section lists whether the Control Station reboot is mandatory. Note that this initial output is long and might scroll off the screen. You will be prompted whether to continue the upgrade after all the upgrade information is displayed.

If you enter no, the upgrade stops, and you are returned to the command prompt.

===========System Information===========

Install Manager Version: 6.0.34-4

Install Manager Command:

/celerra/upgrade/pkg_6.0.34.4/install_mgr

Starts at: Thu Jun 3 10:36:56 2010

From version: 5.6.49-1

To: 6.0.34-4

Dual Control Stations: yes

Slot: slot_0

Blade slot(s): 2, 3, 4, 5

Cabinet family: NS

Cabinet type: NS-HAMMERHEAD

Control Station reboot: multiple (mandatory)

Blade reboot: 1 time (mandatory)

Total number of attempts: 0

=========================================

.

.< other information is posted here>

.

Do you wish to continue with the upgrade now [ yes or no ]? > yes

65



Data Mover rebooting The Data Movers require a reboot in order to run the new version of the VNX for File operating system. When upgrading VNX for File software, if either of the first two digits of the new version number change (for example, 6.0.x to 7.0.x), the Data Mover requires rebooting.

The system information screen is the same as the information in the Control Station rebooting screen.

Limited scheduling of the Data Mover reboot exists. An estimated time for when the Data Mover reboot will happen is shown at the beginning of upgrade:

==============Estimated Time for Services ===============

Current Time: 10:36

Estimated Time when NAS service will be stopped: 10:38

Estimated Time when Data Movers will be reset: 11:28

Estimated Time when NAS service will be restarted: 11:39

Estimated Time when upgrade will be complete: 11:50

==================================================

The Upgrade Data Movers task prompts before it reboots the standby Data Movers. It then provides a menu to reboot the primary Data Movers similar to the following:

==================================================

11:58 [ 46/70 ] Upgrade Data Movers 7 minutes

Start at: Thu Jun 3 11:58:38 2010

All servers must now be rebooted.

Ready to reboot the Standby Data Mover (3). Press the ENTER key to begin...

The following server(s) need to be rebooted to run the new release:

2 ) server_2

Enter slot number(s) or ALL to select server(s) for reboot:

==================================================

When upgrading VNX OE for File, if both of the first two digits of the new version of software remain the same (for example, 7.0.x to 7.0.y), you might delay the rebooting of the Data Movers.

The system information screen is the same as the information in the Control Station rebooting example except for the following:

Data Mover reboot: 1 time (advisory)

66



The Upgrade Data Movers task asks if you wish to reboot the Data Movers. If you enter yes, you will be prompted to press ENTER to reboot the standby Data Movers, then prompted to reboot the primary Data Movers. If you enter no, the Data Movers continue running the current (pre-upgrade) version of VNX software until they are rebooted.

=================================================

12:29 [ 44/67 ] Upgrade Data Movers7 minutes

All servers need to be rebooted to run the new release.

Do you wish to reboot them now? [yes or no]:

Ready to reboot the Standby Data Mover (3). Press the ENTER key to begin...

The following server(s) need to be rebooted to run the new release:

2 ) server_2

Enter slot number(s) or ALL to select server(s) for reboot: Waiting for ping response from server_2 ... :up (0 secs)

=================================================

Schedules during upgrades

As part of the upgrade procedure, schedules that run between midnight and 1:00 A.M. might lose those run times. This affects only schedules that run every X hours of the day (X being from 1 through 12), and not schedules that run at specific hours of the day. This is a known problem and can be corrected by modifying the affected schedules after the upgrade.

Upgrade requirements

Existing read/write-able file system must have 128K minimum free space prior to upgrading to VNX for File version 7.0.

Block OE upgrades

Occasionally, VNX for Block storage connected to a VNX for File requires a software upgrade. When upgrading Block OE software, do not run heavy loads on the VNX for File, as operations may be extremely slow or interrupted.

Block OE software cannot be upgraded unless the Control Station software is upgraded first. Ignoring this rule might make the VNX unmanageable by the VNX for File.

VNX for Block upgrades cannot be made unless the system is running a compatible release of VNX for File software.

67

Troubleshooting and getting help


A warning will appear when you upgrade the VNX for File software if the VNX for Block software needs to be updated. If you receive this warning, finish upgrading the VNX for File software and then upgrade the VNX for Block software as soon as possible to avoid compatibility issues.

Consult the E-Lab Interoperability Navigator at http://Support.EMC.com for all software interdependencies.


Where to get help

EMC support, product, and licensing information can be obtained as follows:

Product information: For documentation, release notes, software updates, or for information about EMC products, licensing, and service, go to EMC Online Support (registration required) at:

http://Support.EMC.com

Troubleshooting: Go to EMC Online Support. After logging in, locate the appropriate Support by Product page.

Technical support: For technical support and service requests, go to EMC Customer Service on EMC Online Support. After logging in, locate the appropriate Support by Product page and choose either Live Chat or Create a service request. To open a service request through EMC Online Support, you must have a valid support agreement. Contact your EMC sales representative for details about obtaining a valid support agreement or to answer any questions about your account.

Note: Do not request a specific support representative unless one has been assigned to your particular system problem.

http://support.emc.com/

https://support.emc.com/

68



Copyright ©2012 EMC Corporation. All Rights Reserved. Published in the USA.

Published December 2012

EMC believes the information in this publication is accurate as of its publication date. The information is subject to change without notice.

THE INFORMATION IN THIS PUBLICATION IS PROVIDED "AS IS." EMC CORPORATION MAKES NO REPRESENTATIONS OR WARRANTIES OF ANY KIND WITH RESPECT TO THE INFORMATION IN THIS PUBLICATION, AND SPECIFICALLY DISCLAIMS IMPLIED WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE.

Use, copying, and distribution of any EMC software described in this publication requires an applicable software license.

EMC2, EMC, and the EMC logo are registered trademarks or trademarks of EMC Corporation in the United States and other countries. All other trademarks used herein are the property of their respective owners.

For the most up-to-date regulatory document for your product line, go to the technical documentation and advisories section on the EMC online support website.

vnx operating environment for file release notes · pdf file03/12/2012 · 1. emc...

Documents