How do I check the output from my job while the job is running?
You can do "checkjob -v $YOUR_JOB_ID" and look for the line following "Allocated Nodes:". The list below it contains the hostname in square brackets with a colon followed by the number of procs you were allocated on that node.
You can ssh into that node from a login node and look in the directory /usr/spool/PBS/spool. It will be /usr/spool/PBS/spool/$JOBID.fslsched.fsl.byu.edu.OU and .ER.
Example:
me@m6int01:~$ checkjob -v 12345 job 12345 (RM job '12345.fslsched.fsl.byu.edu') AName: myjobname State: Running Creds: user:userabcde group:userabcde account:user45678 class:batch qos:normal WallTime: 20:03:33 of 12:12:00:00 SubmitTime: Thu Jul 5 14:57:27 (Time Queued Total: 00:00:44 Eligible: 00:00:43) StartTime: Thu Jul 5 14:58:11 Total Requested Tasks: 1 Req0? TaskCount: 1 Partition: base Memory >= 8192M Disk >= 0 Swap >= 0 Dedicated Resources Per Task: PROCS: 1 MEM: 8192M Utilized Resources Per Task: PROCS: 0.82 MEM: 22G SWAP: 25G Avg Util Resources Per Task: PROCS: 0.82 Max Util Resources Per Task: PROCS: 0.83 MEM: 22G SWAP: 25G Average Utilized Memory: 22726.15 MB Average Utilized Procs: 0.82 TasksPerNode: 1 NodeCount: 1 Allocated Nodes: [m5-4-5:1] Notification Events: JobStart,JobEnd,JobFail Notification Address: email@example.com Task Distribution: m5-4-5 Shell: /bin/bash IWD: /some/path UMask: 0000 OutputFile: m6int02.fsl.byu.edu:/some/path.o12345 ErrorFile: m6int02.fsl.byu.edu:/some/path.e12345 StartCount: 1 Partition List: [ALL] SrcRM: base DstRM: base DstRMJID: 12345.fslsched.fsl.byu.edu Submit Args: file.sh Flags: BACKFILL,PREEMPTOR,FSVIOLATION Attr: BACKFILL,FSVIOLATION,checkpoint StartPriority: 19873 PE: 2.78 Reservation '12345' (-20:03:51 -> 11:15:56:09 Duration: 12:12:00:00)me@m6int01:~$ ssh m5-4-5 me@m5-4-5:~$ cd /usr/spool/PBS/spool me@m5-4-5:/usr/spool/PBS/spool$ ls 12345.fslsched.fsl.byu.edu.ER 12345.fslsched.fsl.byu.edu.OU me@m5-4-5:/usr/spool/PBS/spool$ tail 12345.fslsched.fsl.byu.edu.OU some of the last lines of your really important job's output are now shown me@m5-4-5:/usr/spool/PBS/spool$
In this case the hostname is m5-4-5.
Last changed on Fri Jul 6 11:45:22 2012