Help:HPCC Torque
Submitting Jobs to the CSDMS HPCC
The CSDMS High Performance Computing Cluster uses Torque/Maui as a job scheduler. With Torque you can allocate resources, schedule and manage job execution, monitor and view the status of your jobs.
Torque uses instructions given on the command line and embedded within comments of the shell script that runs your program. This page describes basic Torque usage. Please visit the Torque website for a more complete guide.
The Torque Cheat Sheet
Frequently Used Commands
Command | Description | Basic Usage | Example |
---|---|---|---|
qsub | Submit a pbs job | qsub [script] | > qsub job.pbs |
qstat | Show status of pbs batch jobs | qstat [job_id] | > qstat 44 |
qdel | Delete pbs batch job | qdel [job_id] | > qdel 44 |
qhold | Hold pbs batch jobs | qhold [job_id] | > qhold 44 |
qrls | Release hold on pbs batch jobs | qrls [job_id] | > qrls 44 |
Check Queue and Job Status
Command | Description |
---|---|
qstat -q | List all queues |
qstat -a | List all jobs |
qstat -au <userid> | list jobs for userid |
qstat -r | List running jobs |
qstat -f <job_id> | List full information about job_id |
qstat -Qf <queue> | List full information about queue |
qstat -B | List summary status of the job server |
pbsnodes | List status of all compute nodes |
Within-Script Torque Commands
Command | Description |
---|---|
#PBS -N myjob | Set the job name |
#PBS -m ae | Mail status when the job completes |
#PBS -M your@email.address | Mail to this address |
#PBS -l nodes=4 | Allocate specified number of nodes |
#PBS -l walltime=1:00:00 | Inform the PBS scheduler of the expected runtime |
Basic Usage
Torque is able to dynamically allocate resources for your job. You simply need to submit it and it will find the processors for you.
Submit a job: To submit a job you must write a shell script that torque will run. The idea is to create a Torque job file and then run "qsub job_file". Torque will run the job file with the options specified.
A basic job script example: <geshi> > cat test.sh
#!/bin/sh
date sleep 30
</geshi>
This script will run date, and then sleep for 20 seconds on one processor. This script is not really useful until you replace the sleep and date commands with process intensive commands.
To submit the job to the queue, use qsub.
<geshi> >qsub test.sh 2607.servername.colorado.edu </geshi>
This job has the id 2607 and will be submitted to the default queue. When there is enough resources available, this job will run. On completion, Standard output and Standard error are saved into files where you ran the program from.
<geshi> > ls test.sh.o2607 test.sh.e2607 </geshi>
Delete a job: To delete a job use "qdel". Qdel will remove the job from the queue, and it will not be run. If it is being run it will stop the job.
<geshi> >qdel 2607 </geshi>
Check the status of a job: To check the status of a job use "qstat". Qstat is a command that will return all queued and running jobs.
<geshi> >qstat
Job id Name User Time Use S Queue ------------------------- ---------------- --------------- -------- - ----- 2607.servername STDIN username 0 R workq 2608.servername STDIN username 0 Q workq
</geshi>
The 'S' parameter tells the status of the job. 'R' for running, 'Q' for queued.
Advanced usage
Torque allows you to use advanced features and customizations when running jobs. The below sections are continuations of the sections above.
Job Submission: There are options in the shell script that can be used to customize your job.
A Basic script. <geshi> >cat test.sh
#!/bin/bash #PBS -N testjob
cat $PBS_NODEFILE sleep 30
</geshi>
$PBS_NODEFILE is the location of a file that contains a list of the nodes allocated for this job.
- PBS specifies an option to Torque. There are many listed below, but
more can be found in the man page for qsub.
<geshi>
- PBS -r n # The job is not rerunnable.
- PBS -r y # The job is rerunnable
- PBS -q testq # The queue to submit to
- PBS -N testjob # The name of the job
- PBS -o testjob.out # The file to print the output to
- PBS -e testjob.err # The file to print the error to
- Mail Directives
- PBS -m abe # The points durring the execution to send an email
- PBS -M me@colorado.edu # Who to Mail to
- PBS -l walltime=01:00:00 # Specify the walltime
- PBS -l pmem=100mb # Memory Allocation for the Job
- PBS -l nodes=4 # Number of nodes to Allocate
- PBS -l nodes=4:ppn=3 # Number of nodes and the number processors per node
</geshi>
You can use any of the above options in the script to customize your job. If all of the above options are used, the job will be named testjob and be put into the testq. It will only run for 1 hour and mail me@colorado.edu at the beginning and end of the job. It will use 4 nodes with 3 processors per node, with a total of 12 processors and 100 mb of memory.
Check the status of a job: Torque and Maui allow you to check the status of jobs and the queue status.
In Torque: Qstat has many options for checking a job status. The basic way is running the command with out any options which is showed above. Again the man pages are the best resources for information.
Other options include: -n, -f, -Q, -B, -u, -q
The -n option will show which nodes are running which jobs. <geshi> >qstat -n
server.colorado.edu: Req'd Req'd Elap Job ID Username Queue Jobname SessID NDS TSK Memory Time S Time -------------------- -------- -------- ---------------- ------ ----- --- ------ ----- - ----- 78.server.colorado user workq STDIN 4811 -- -- -- -- R -- node34/0 79.server.colorado user workq STDIN 4830 -- -- -- -- R -- node34/1 80.server.colorado user workq STDIN 3867 -- -- -- -- R -- node33/0 81.server.colorado user workq STDIN 4821 -- -- -- -- R -- node32/0 82.server.colorado user workq STDIN 4840 -- -- -- -- R -- node32/1 83.server.colorado user workq STDIN 4859 -- -- -- -- R -- node32/2
</geshi>
The -f option will show the full details for a specified job. <geshi> >qstat -f 78 Job Id: 84.server.colorado.edu
Job_Name = STDIN Job_Owner = username@server.colorado.edu resources_used.cput = 00:00:00 resources_used.mem = 1704kb resources_used.vmem = 8028kb resources_used.walltime = 00:00:01 job_state = R queue = workq server = server.colorado.edu Checkpoint = u ctime = Fri Apr 24 16:21:51 2009 Error_Path = server.colorado.edu:/tmp/STDIN.e84 exec_host = node34/0 Hold_Types = n Join_Path = n Keep_Files = n Mail_Points = a mtime = Fri Apr 24 16:21:53 2009 Output_Path = server.colorado.edu:/tmp/STDIN.o84 Priority = 0 qtime = Fri Apr 24 16:21:51 2009 Rerunable = True Resource_List.neednodes = node34 session_id = 4877 substate = 42 Variable_List = PBS_O_HOME=/tmp,PBS_O_LOGNAME=username, PBS_O_PATH= /usr/local/bin:/usr/bin PBS_O_SHELL=/bin/tcsh,PBS_SERVER=server.colorado.edu, PBS_O_HOST=server.colorado.edu,PBS_O_WORKDIR=/tmp/ PBS_O_QUEUE=workq euser = username egroup = server hashname = 84.server.colorado.edu queue_rank = 83 queue_type = E etime = Fri Apr 24 16:21:51 2009 start_time = Fri Apr 24 16:21:53 2009 start_count = 1
</geshi>
The -u option will show all jobs owned the specified user.
The -Q option will show the queue information. If a specific queue is specified it will only show the information from that queue.
<geshi> >qstat -Q
Queue Max Tot Ena Str Que Run Hld Wat Trn Ext T ---------------- --- --- --- --- --- --- --- --- --- --- - testing 0 0 yes yes 0 0 0 0 0 0 E normal 8 1 yes yes 0 1 0 0 0 0 E short 0 0 yes yes 0 0 0 0 0 0 E long 0 3 yes yes 0 3 0 0 0 0 E special 0 0 yes yes 0 0 0 0 0 0 E
</geshi>
Maui
If maui is installed on your system, you will have access to another set of tools. One of these is showq. showq is a tool like qstat. It will show the queue information.
<geshi> >showq
ACTIVE JOBS-------------------- JOBNAME USERNAME STATE PROC REMAINING STARTTIME
624 user1 Running 4 21:00:01 Fri Apr 24 13:34:17 621 user2 Running 2 95:21:19:49 Mon Apr 20 13:54:06 622 user2 Running 2 95:21:23:06 Mon Apr 20 13:57:23 623 user2 Running 2 96:04:13:37 Mon Apr 20 20:47:54
4 Active Jobs 10 of 20 Processors Active (50.00%) 5 of 7 Nodes Active (71.43%)
IDLE JOBS---------------------- JOBNAME USERNAME STATE PROC WCLIMIT QUEUETIME
0 Idle Jobs
BLOCKED JOBS---------------- JOBNAME USERNAME STATE PROC WCLIMIT QUEUETIME
Total Jobs: 4 Active Jobs: 4 Idle Jobs: 0 Blocked Jobs: 0
</geshi>