Accounting and Job History Retrieval

Slurms job records can show information including memory usage, time start and end, and other details about your jobs submitted to the cluster. This can be helpful for creating a more efficient resource allocation to your jobs or help with looking into any errors you may run into.


How to show job information on a specific job:

hpc@koko-login:$ sacct -j jobid

You can also add a format to the sacct command:

hpc@koko-login:$ sacct -j jobid –format=User,Jobid,Jobname,partition,state,time,start,end,elapsed,nodelist

To look at jobs starting from a specific date:

hpc@koko-login:$ sacct –starttime 2021-02-18 –format=User,Jobid,Jobname,partition,state,time,start,end,elapsed,nodelist

A list of available format parameters and other options for sacct can be found here


How to look at jobs currently in queue:

hpc@koko-login:$ squeue -u netid

A list of parameters for squeue can be found here


How to look at available nodes:

hpc@koko-login:$ sinfo

Additional information on sinfo can be found here




Posted in HPC