Useful b-commands¶
The command-line programs that you use to interact with LSF start with the letter “b”: bsub, bqueues, bkill, etc.
Hence, we refer to this group of commands as b-commands. This page lists some of the most useful b-commands.
For the complete documentation of all the b-commands, see the LSF Command Reference.
bjobsLists your active jobs. For example
[liam.bindle@compute1-client-1 ~]$ bjobs JOBID USER STAT QUEUE FROM_HOST EXEC_HOST JOB_NAME SUBMIT_TIME 59060 liam.bi RUN general-in compute1-cl 8*compute1- *#zsh;bash Feb 20 09:56
Alternatively, you can use the
-l <JOB_NUMBER>option for more verbose output[liam.bindle@compute1-client-1 ~]$ bjobs -l 15691 Job 15691, User liam.bindle, Project default, Application docker1, Job Group /liam.bindle/default, User Group , Status PEND, Queue rvmartin-interactive, Interactiv e pseudo-terminal shell mode, Command <#!/bin/bash;#BSUB - Is;#BSUB -n 24;#BSUB -R "rusage[mem=200000] span[hosts=1]" ;#BSUB -q rvmartin-interactive;#j#BSUB -a 'docker(liambind le/mypyenv)';##BSUB -a 'docker(registry.gsc.wustl.edu/sleo ng/base-engineering)';#BSUB -a 'docker(registry.gsc.wustl. edu/sleong/base-engineering-gcc)'; #source /init.rc;source $HOME/compute1/compute-node.rc;cd /my-projects;zsh;#bash> , Esub registry.gsc.wustl.edu/sleong/base-engineer ing-gcc)> Wed Jan 15 08:29:47: Submitted from host compute1-client-1.ris.wustl.edu>, CWD $HOME, 24 Task(s), Requested Resources ; PENDING REASONS: Not enough job slot(s): 25 hosts; RUNLIMIT 1440.0 min MEMLIMIT 195.3 G SCHEDULING PARAMETERS: r15s r1m r15m ut pg io ls it tmp swp mem loadSched - - - - - - - - - - - loadStop - - - - - - - - - - - RESOURCE REQUIREMENT DETAILS: Combined: select[(defined(docker)) && (type == any)] order[r15s:pg] rusage[mem =200000.00] span[hosts=1] Effective: -bkillKill one or more jobs. Pass this command the jobs that you want to kill. For example
[liam.bindle@compute1-client-1 ~]$ bkill 59060 Job <59060> is being terminated
bqueuesCheck how much activity there is on the queues. For example,
[liam.bindle@compute1-client-1 ~]$ bqueues QUEUE_NAME PRIO STATUS MAX JL/U JL/P JL/H NJOBS PEND RUN SUSP datatransfer 10 Open:Active - - - - 0 0 0 0 general 10 Open:Active - - - - 3984 111 3873 0 general-interac 10 Open:Active - - - - 3 0 3 0 fitzp 10 Open:Active - - - - 0 0 0 0 fitzp-interacti 10 Open:Active - - - - 0 0 0 0 pathology 10 Open:Active - - - - 0 0 0 0 pathology-inter 10 Open:Active - - - - 0 0 0 0 rvmartin 10 Open:Active - - - - 0 0 0 0 rvmartin-intera 10 Open:Active - - - - 0 0 0 0 tychele 10 Open:Active - - - - 0 0 0 0 tychele-interac 10 Open:Active - - - - 0 0 0 0 irahall 10 Open:Active - - - - 0 0 0 0 irahall-interac 10 Open:Active - - - - 16 0 16 0 timley 10 Open:Active - - - - 0 0 0 0 timley-interact 10 Open:Active - - - - 0 0 0 0
bhostsChecks how much activity is on each exec node. Pass this command the queue that you want to check.
[liam.bindle@compute1-client-1 ~]$ bhosts rvmartin HOST_NAME STATUS JL/U MAX NJOBS RUN SSUSP USUSP RSV compute1-exec-10.r ok - 36 0 0 0 0 0 compute1-exec-11.r ok - 36 0 0 0 0 0 compute1-exec-12.r ok - 36 0 0 0 0 0 compute1-exec-13.r ok - 36 0 0 0 0 0 compute1-exec-14.r ok - 36 0 0 0 0 0 compute1-exec-15.r ok - 36 0 0 0 0 0 compute1-exec-16.r ok - 36 0 0 0 0 0 compute1-exec-17.r ok - 36 0 0 0 0 0 compute1-exec-18.r ok - 36 0 0 0 0 0 compute1-exec-19.r ok - 36 0 0 0 0 0 compute1-exec-20.r ok - 36 0 0 0 0 0 compute1-exec-21.r ok - 36 0 0 0 0 0 compute1-exec-22.r closed - 36 0 0 0 0 0 compute1-exec-23.r ok - 36 0 0 0 0 0 compute1-exec-24.r ok - 36 0 0 0 0 0 compute1-exec-25.r ok - 36 0 0 0 0 0 compute1-exec-26.r ok - 36 0 0 0 0 0 compute1-exec-27.r ok - 36 0 0 0 0 0 compute1-exec-28.r ok - 36 0 0 0 0 0 compute1-exec-29.r ok - 36 0 0 0 0 0 compute1-exec-3.ri closed - 1 0 0 0 0 0 compute1-exec-30.r ok - 36 0 0 0 0 0 compute1-exec-4.ri ok - 36 0 0 0 0 0 compute1-exec-5.ri ok - 36 0 0 0 0 0 compute1-exec-6.ri ok - 36 0 0 0 0 0 compute1-exec-7.ri ok - 36 0 0 0 0 0 compute1-exec-8.ri closed - 36 0 0 0 0 0 compute1-exec-9.ri ok - 36 0 0 0 0 0
bgaddCreates a job group. For example
[liam.bindle@compute1-client-1 ~]$ bgadd -L 30 /liam.bindle/test_group Job group was added.
In your job script, you specify the job’s group with the
-goption. For example, to include a job in this group you would have the following line in your job script#BSUB -g /liam.bindle/test_group