HPC简单作业测试

HPC简单作业测试
创建脚本1.sh#!/bin/bash# 1.sh - 从0到1亿的死循环每秒钟写一个数字到0.tmp覆盖写入# 循环从0到100000000for((i0;i100000000;i));doecho$i0.tmpsleep1done2.sh#!/bin/bash#JSUB -q gpu1#JSUB -m gpu02#JSUB -n 1#JSUB -o output.%J#JSUB -e error.%J#JSUB -J write_file_job# 切换到当前目录并执行1.shcd/public/home/10201401498bash1.sh创建作业-bash-4.2$ jsub2.sh Job24785is submitted to queuegpu1and default applicationdefault检查作业查看作业状态-bash-4.2$ jjobs JOBIDUSERSTAT QUEUE FROM_HOST EXEC_HOST JOB_NAME SUBMIT_TIME ORDER247851020140* RUN gpu1 login gpu02 *te_file_job Jul 0817:33 -查看作业详细信息-bash-4.2$ jjobs-l24785Job24785, Job Namewrite_file_job, User10201401498, Projectdefault, StatusRUN, Queuegpu1, Applicationdefault, APS Pri ority5000, Command#!/bin/bash;#JSUB -q gpu1;#JSUB -mgpu02;#JSUB -n 1;#JSUB -o output.%J;#JSUB -e error.%J;#JSUB-Jwrite_file_job;# 切换到当前目录并执行1.sh;cd /public/home/10201401498;bash1.shWed Jul 0817:33:35: Submitted fromhostlogin, CWD/public/home/10201401498, OutFile Fileoutput.24785, Error Fileerror.24785, Specified Hostsgpu02,1Processors Requested;Wed Jul 0817:33:36: Started ongpu02, Execution Home/public/home/10201401498, Execution CWD/public/home/10201401498, Execution U ser10201401498;Wed Jul 0817:34:38: Resource usage collected. The CPUtimeused is0.02seconds. IDLE_RATE(cputime/slots/runtime):0.0004CPU(s)Utilization:0.02% MEM:22Mbytes;SWAP:0Bbytes PGIDS:15200;PIDS:1520015487152281520615229RUNLIMIT2880.0min SCHEDULING PARAMETERS: r15s r1m r5m r15m ut pg iolsLoadSched - - - - - - - - LoadStop - - - - - - - - it tmp swap mem LoadSched - - - - LoadStop - - - -检查文件是否生成查看 0.tmp 文件是否存在-bash-4.2$ls-la0.tmp -rw-r--r--110201401498jhdomainusers4Jul817:360.tmp如果文件存在查看内容-bash-4.2$cat0.tmp212查看输出和错误日志-bash-4.2$catoutput.24785查看错误输出-bash-4.2$caterror.24785实时监控作业查看作业动态输出如果有-bash-4.2$ jctrl peek24785output from stdoutoutput from stderr实时查看日志文件-bash-4.2$tail-foutput.24785 error.24785output.24785error.24785终止作业终止作业-bash-4.2$ jctrlkill24785Job24785is being terminated强制终止如果上面命令无效jctrlkill-f24785确认作业已终止查看当前作业状态-bash-4.2$ jjobs No unfinished job found查看所有作业包括已完成的-bash-4.2$ jjobs-aJOBIDUSERSTAT QUEUE FROM_HOST EXEC_HOST JOB_NAME SUBMIT_TIME ORDER247851020140* EXIT gpu1 login gpu02 *te_file_job Jul 0817:33 -清理作业记录清理已完成作业-bash-4.2$ jctrl clean24785Job24785is being cleaned