+ unobtrusive power proportionality for torque: design and implementation arka bhattacharya...

+Unobtrusive power proportionality for Torque: Design and Implementation

Arka BhattacharyaAcknowledgements:Jeff Anderson LeeAndrew KrioukovAlbert Goto

+Introduction

What is power proportionality ? Performance-power ratio at all performance levels is

equivalent to that at the maximum performance level Servers consume a high percentage of their max power

even idle Hence, power proportionality => switch off idle servers

+NapSAC – Krioukov et.al.

3

IPSRequests

Power

Computational “Spinning Reserve”

Load DistributionScheduling

Power ManagementWikiPedia Request Rate

4/12/11CPS 2011

+The need for power proportionality of IT equipment in Soda Hall

Soda Hall Power : 450-500kW

Cluster Room Power: 120-130kW (~25%)

Total HVAC for cluster rooms : 75-85kW(~15%)

+PSI Cluster

Cluster Room Power: 120-130kW (~25% of Soda)

PSI Cluster: 20-25kW (~5% of Soda)

Total HVAC for cluster rooms : 75-85kW(~15% of Soda)

Total HVAC for PSI Cluster room : 20-25kW(~5% of Soda)

+The PSI Cluster

PSI Cluster Consumes ~20-25kW of power irrespective of workload. Contains about 110 servers.

Recently server faults have reduced the size of the cluster to 78 servers. (The faulty servers mostly are powered on all the time)

Used mainly by NLP, Vision, AI and ML graduate students.

It is an HPC Cluster running Torque

+PSI Cluster

+Possible Energy savings

Can save ~ 50% of the energy

+Current state :

+ Result:

10 kW

We save 49% of the energy

+What is Torque?

Tera-scale Open-source Research and QUEue manager

Built upon original Portable Batch System (PBS) project

Resource manager: Manages availability of, and requests for, compute node resources

Used by most academic institutions throughout the world for batch processing.

+Maui Scheduler

Job scheduler

Implements and manages:

Scheduling policies

Dynamic priorities

Reservations

Fairshare

+Sample Job Flow

Script submitted to TORQUE specifying required resources

Maui periodically retrieves from TORQUE list of potential jobs, available node resources, etc.

When resources become available, Maui tells TORQUE to execute certain jobs on particular nodes

TORQUE dispatches jobs to the PBS MOMs (machine oriented miniserver) running on the compute nodes - pbs_mom is the process that starts the job script

Job status changes reported back to Maui, information updated

+Why are we building power-proportional Torque ?

To shed load in Soda Hall

To investigate why production clusters don’t implement power proportionality

To integrate power-proportionality into a software used in many clusters throughout the world

+Desirables from an unobtrusive power proportionality feature

Avoid modifications to torque source code

Only use existing torque interfaces

Make the feature completely transparent to end users

Maintain system responsiveness

Centralized

No dependence resource manager/scheduler version

+Analysis of the psi cluster

Logs : Active and Idle Queue Log Job placement statistics

Logs exist for 68 days in Feb-April,2011

Logs were recorded once every minute

Logs contain information of ~169k jobs , ~40 users

+Type of servers in the psi cluster

Server Make Number of Cores

Memory Count

Dell 2 3GB 64

Dell 8 16GB 21

Intel 8 48GB 28

Intel 24 256GB 4

Total : 117

• Each server class is further divided according to various features

• Not all servers listed above are switched on all the time

+CDF of server idle duration

TAKEAWAY 1: Most idle periods are small

+Contribution of server idle period to total

TAKEAWAY 2: To save energy, tackle the large idle periods

+CDF of job durations

(50,500s)

BATCH

INTERACTIVE

TAKEAWAY 3: Most jobs are long. Hence slight increase in queuing time wont hurt

+Summary of takeaways

Small server idle times, though numerous, contribute very less to total server idle time.

Power proportionality algorithm need not be aggressive in switching of servers

Waking servers takes 5 min. Considered to the running time of a job, it is negligible

+Loiter Time vs Energy Savings

+

Design of unobtrusive Power Proportionality for Torque

+Using Torque interfaces

What useful state information does torque/maui maintain ?

Maintains the state(active/offline/down) of each server, and jobs running on it. Obtained through “pbsnodes” command

Maintains a list of running and queued jobs Obtained through “qstat” command

Maintains job constraints and scheduling details of each job Obtained through “checkjob” command

+First implementation- State machine for each server

Active

Offline

Down

Waking

• Server_idle_time > LOITER_TIME

• Server_offline_time >OFFLINE_LOITER_TIME

• No job has been scheduled on server

• Idle job exists

• Server has woken up

Problematic Server

• Server not waking

• If idle job can be scheduled on server

+Does not work !

Each job is submitted to a specific queue, Must ensure right server wakes up.

+Next implementation-State machine for each server

Active

Offline

Down

Waking

• Server_idle_time > LOITER_TIME

• Server_offline_time > OFFLINE_LOITER_TIME

• No job has been scheduled on server

• Idle job exists• Server

belongs to desired queue

• Server has woken up

Problematic Server

• Server not waking

• If idle job can be scheduled on server

+Still did not work !

Each job has specific constraints which torque takes into account while scheduling

Job constraints can be obtained through “checkjob” command.

+Next implementation-State machine for each server

Active

Offline

Down

Waking

• Server_idle_time > LOITER_TIME

• Server_offline_time > OFFLINE_LOITER_TIME

• No job has been scheduled on server

• Idle job exists• Server belongs

to desired queue

• Server satisfies job constraints

• Server has woken up

Problematic Server

• Server not waking

• If idle job can be scheduled on server

+Scheduling problem: Job submission characteristics

Users tend to submit multiple jobs at a time (often >20)

Torque has its own fairness mechanisms, which wont schedule all the jobs even if there are free servers.

To accurately predict which jobs Torque will schedule, and not to switch on extra servers, we should emulate the Torque scheduling logic !

Ties Power Proportionality feature to specific Torque Policy

Solution : Switch on only a few servers at a time to check if torque schedules the idle job

+Next implementation-State machine for each server

Active

Offline

Down

Waking

• Server_idle_time > LOITER_TIME

• Server_offline_time > OFFLINE_LOITER_TIME

• No job has been scheduled on server

• Idle job exists• Server belongs to

desired queue• Server satisfies job

constraints• Switch on only a few

servers at a time

• Server has woken up

Problematic Server

• Server not waking

• If idle job can be scheduled on server

+Maintain responsiveness/headroom

The Debug cycle usually contains the users running short jobs and validating the output

If no server satisfying job contraints are switched on, a user might have to wait a long time to validate if his job is running

If jobs throw errors, he might have to wait for an entire server power cycle to run his modified job

Solution : Group servers according to features. In each group, have a limited numbers of servers as

spinning reserve all the time

+Final implementation-State machine for each server

Active

Offline

Down

Waking

• Server_idle_time > LOITER_TIME

• Server_offline_time >OFFLINE_LOITER_TIME

• No job has been scheduled on server

• Switching off servers leaves no headroom

• Idle job exists• Server belongs to desired queue• Server satisfies job constraints• Switch on only MAX_SERVERS at a

time• Switch on server to maintain

headroom

• Server has woken up

Problematic Server

• Server not waking

• If idle job can be scheduled on server

+But the servers don’t wake up !!!

Each server has to bootstrap a list of service, such as network file systems, work directories, portmapper, etc

Often these bootstraps fail, and hence servers are left in an undesired state ( e.g with no home directories mounted to write user output to ! )

Solution : Have a health-check script on each server Check for proper configurations of useful services, and

make server available for scheduling only if health-check succeeds.

+Power Proportional Torque at a glance:

Completely transparent to user

Did not modify torque source code

1000 line python script which runs only on torque master server

Halts servers through ssh

Wake servers through wake-on-lan

Separates scheduling policy from mechanism. It allows torque to dictate the scheduling policy.

+Deployment

Deployed on 57 of the 78 active nodes in the psi cluster. Total number of cores = 150

Servers were classified into 5 groups based on features.

HEADROOM_PER_GROUP = 3

MAX_SERVERS_TO_WAKE_AT_A_TIME = 5

LOITER_TIME = 7 minutes

OFFLINE_LOITER_TIME = 3 minutes

+Average Statistics

Deployed since last week

~800 jobs analyzed

Avg utilization of cluster = 40%

% Energy saved = 49%

+

+Results:

+HVAC power savings

+Number of servers powered on at a time:

Headroom

+Expected vs Actual savings

+Submission vs Execution profile

5/17/12 12:00 5/18/12 0:00 5/18/12 12:00 5/19/12 0:00 5/19/12 12:00 5/20/12 0:00 5/20/12 12:00 5/21/12 0:000

20

40

60

80

100

120Submission Profile Execution profile

Time

Nu

mb

er

of

Act

ive C

ore

s

+CDF of job queue time as a percentage of job length

+Conclusions – what we achieved

Power proportionality is easy to achieve for torque without changing any source code at all

The script could be run on any standard torque cluster to save energy.

Switching servers back on in a consistent state is the single biggest roadblock to deployment of script.

We saved a max of ~17kW of power is Soda Hall (~3%). This was only half the psi cluster !