[slurm-users] howto list/get all scripts run by a job?

[mercan]

Hi;

For running jobs, you can get the running script with using:

scontrol write batch_script  "$SLURM_JOBID" -

command. the - parameter reqired for screen output.

Ahmet M.


On 19.06.2020 12:25, Adrian Sevcenco wrote:
> On 6/18/20 9:35 AM, Loris Bennett wrote:
>> Hi Adrain,
> Hi
>
>> Adrian Sevcenco <Adrian.Sevcenco at spacescience.ro> writes:
>>
>>> Hi! I'm trying to retrieve the actual executable of jobs but i did 
>>> not find how
>>> to do it .. i would like to found this for both case when the job is 
>>> started
>>> with sbatch or with srun.
>>
>> For running jobs:
>>
>>    scontrol show job <job id>
> well, this was the first thing i tried but i have a null command
>
> [root at alien ~]# scontrol show job 2794270
> JobId=2794270 JobName=AliEn.4865.575
>    UserId=aliprod(1000) GroupId=aliprod(1000) MCS_label=N/A
>    Priority=13338 Nice=0 Account=aliprod QOS=normal WCKey=*
>    JobState=RUNNING Reason=None Dependency=(null)
>    Requeue=0 Restarts=0 BatchFlag=1 Reboot=0 ExitCode=0:0
>    RunTime=09:20:37 TimeLimit=1-00:00:00 TimeMin=N/A
>    SubmitTime=2020-06-19T02:22:45 EligibleTime=2020-06-19T02:22:45
>    AccrueTime=2020-06-19T02:22:45
>    StartTime=2020-06-19T02:32:10 EndTime=2020-06-20T02:32:10 Deadline=N/A
>    SuspendTime=None SecsPreSuspend=0 LastSchedEval=2020-06-19T02:32:10
>    Partition=alien AllocNode:Sid=alien.spacescience.ro:4865
>    ReqNodeList=(null) ExcNodeList=(null)
>    NodeList=alien-0-62
>    BatchHost=alien-0-62
>    NumNodes=1 NumCPUs=1 NumTasks=1 CPUs/Task=1 ReqB:S:C:T=0:0:*:*
>    TRES=cpu=1,mem=2600M,node=1,billing=1
>    Socks/Node=* NtasksPerN:B:S:C=0:0:*:* CoreSpec=*
>    MinCPUsNode=1 MinMemoryCPU=2600M MinTmpDiskNode=0
>    Features=(null) DelayBoot=00:00:00
>    OverSubscribe=OK Contiguous=0 Licenses=(null) Network=(null)
>    Command=(null)
>    WorkDir=/tmp
>    StdErr=/dev/null
>    StdIn=/dev/null
>    StdOut=/dev/null
>    Power=
>
> this is slurm 19.05.2 and my purpose is only for running jobs
>
> moreover it's not clear to me what are the steps of job submission and 
> the processes involved ..
>
> it seems to be that slurmstepd (with identification of jobid.batch) 
> start a slurm_script that i think that is the actually submitted script
>
> the job start a srun with a script (specified in submiited script)
> and at the moment i get an slurmstepd[jobid.0] where this is run
>
> so, at this moment it would be enough if given a job id i can get the 
> submission script...
>
> is there a way to do it? (beside get the node from squeue and then ssh 
> on node and ps grep?)
>
> Thank you!
> Adrian
>
>
>>
>> For completed jobs the information about the executable is not kept by
>> the standard accounting mechanism.  However, it is possible to extract
>> more information yourself from either the prolog or epilog and save this
>> somewhere.
>>
>> Cheers,
>>
>> Loris
>>
>