Translating Slurm and LSF-HPC JOBIDs

Example 9-7 Using the bhist Command (Long Output)

$ bhist -l 24

Job <24>, User <lsfadmin>, Project <default>, Interactive pseudo-terminal shell mode, Extsched <SLURM[nodes=4]>, Command </bin/bash> date and time stamp: Submitted from host <n2>, to Queue <normal>, CWD <$HOME>,

4 Processors Requested, Requested Resources <type=any>;

date and time stamp: Dispatched to 4 Hosts/Processors <4*lsfhost.localdomain>;

date and time stamp: slurm_id=22;ncpus=8;slurm_alloc=n[5-8];

date and time stamp: Starting (Pid 4785);

Summary of time in seconds spent in various					states by
date and time stamp
PEND	PSUSP	RUN	USUSP	SSUSP	UNKWN	TOTAL
11	0	124	0	0	0	135

LSF-HPC and SLURM are independent resource management components of the HP XC system. They maintain their own job identifiers (JOBIDs). It may be useful to be able to determine which the SLURM_JOBID environment variable matches an LSF JOBID, and vice versa.

When a job is submitted to LSF-HPC, it is given an LSF JOBID, as in this example:

$ bsub -o %J.out -n 8 sleep 300

Job <99> is submitted to default queue <normal>

The following is the sequence of events when a SLURM JOBID is assigned:.

•No SLURM_JOBID exists while the job is PENDing in LSF-HPC.

•After LSF-HPC determines that the resources are available in SLURM for this job, LSF-HPC requests an allocation in SLURM.

•After the SLURM allocation is established, there is a corresponding SLURM JOBID for the LSF JOBID. Use the bjobs command to view the SLURM JOBID:

$ bjobs -l 99 grep slurm

date and time stamp: slurm_id=123;ncpus=8;slurm_alloc=n[13-16];

The SLURM JOBID is 123 for the LSF JOBID 99.

You can also find the allocation information in the output of the bhist command:

$ bhist -l 99 grep slurm

date and time stamp: slurm_id=123;ncpus=8;slurm_alloc=n[13-16];

When LSF-HPC creates an allocation in SLURM, it constructs a name for the allocation by combining the LSF cluster name with the LSF-HPC JOBID. You can see this name with the scontrol and sacct commands while the job is running:

$ scontrol show job grep Name

Name=hptclsf@99

$ sacct -j 123
Jobstep	Jobname	Partition	Ncpus	Status	Error
----------	------------------	---------- -------		----------	-----
123	hptclsf@99	lsf	8	RUNNING	0
123.0	hptclsf@99	lsf	0	RUNNING	0

In these examples, the jobname is hptclsf@99; the LSF job ID is 99.

Note that the scontrol show job command keeps jobs briefly after they finish, then it purges itself; this is similar with the bjobs command. The sacct command continues to provide job information after the job has finished; this is similar to bhist command:

Using LSF-HPC 83

$ bhist -l 24

date and time stamp

PEND