2009-10-15

• ​Summit'09 - OGF 27

Torque

• 撰寫 pbs script

  jazz@bio037:~$ cat myscript
  #!/bin/bash
  ### Job 名稱
  #PBS -N mytest
  ### 輸出檔案
  #PBS -e /home/jazz/mytest.err
  #PBS -o /home/jazz/mytest.log
  ###================================================
  # 顯示目錄及時間資訊
  echo Working directory is $PBS_O_WORKDIR
  cd $PBS_O_WORKDIR
  echo Running on host `hostname`
  echo Time is `date`
  echo Directory is `pwd`
  # 執行檔案
  date
  

• 丟 job

  jazz@bio037:~$ qsub < myscript
  30.bio037
  

• 查 job 執行過程

  jazz@bio037:~$ tracejob 30
  /var/spool/torque/mom_logs/20091015: No matching job records located
  
  Job: 30.bio037
  
  10/15/2009 00:38:59  S    enqueuing into batch, state 1 hop 1
  10/15/2009 00:38:59  S    Job Queued at request of jazz@bio037, owner = jazz@bio037, job name = mytest, queue =
                            batch
  10/15/2009 00:38:59  S    Job Modified at request of Scheduler@bio037
  10/15/2009 00:38:59  S    Exit_status=0 resources_used.cput=00:00:00 resources_used.mem=0kb
                            resources_used.vmem=0kb resources_used.walltime=00:00:00
  10/15/2009 00:38:59  L    Job Run
  10/15/2009 00:38:59  S    Job Run at request of Scheduler@bio037
  10/15/2009 00:38:59  A    queue=batch
  10/15/2009 00:38:59  A    user=jazz group=jazz jobname=mytest queue=batch ctime=1255538339 qtime=1255538339
                            etime=1255538339 start=1255538339 owner=jazz@bio037 exec_host=bio013/0
                            Resource_List.neednodes=1 Resource_List.nodect=1 Resource_List.nodes=1
  10/15/2009 00:38:59  A    user=jazz group=jazz jobname=mytest queue=batch ctime=1255538339 qtime=1255538339
                            etime=1255538339 start=1255538339 owner=jazz@bio037 exec_host=bio013/0
                            Resource_List.neednodes=1 Resource_List.nodect=1 Resource_List.nodes=1 session=5366
                            end=1255538339 Exit_status=0 resources_used.cput=00:00:00 resources_used.mem=0kb
                            resources_used.vmem=0kb resources_used.walltime=00:00:00
  10/15/2009 00:39:07  S    Post job file processing error
  10/15/2009 00:39:07  S    dequeuing from batch, state COMPLETE
  

• 每個 Job 都可以用 jobid 去查執行的 host 是哪些，在 exec_host 這個變數

                            etime=1255538339 start=1255538339 owner=jazz@bio037 exec_host=bio013/0
  

• 從錯誤訊息，可以明白每台 pbs_mom 執行過的 job 都會紀錄在 /var/spool/torque/mom_logs/日期

  /var/spool/torque/mom_logs/20091015