How can I see who is hogging all of the resources on sungrid engine?

106 Views Asked by At

At my job we using sungrid qstat, qsub, etc.

Is there a way to see the percentage of resources currently used by each user? I know there is qhost -u "*" but this is a bit more difficult to interpret b/c it doesn't show how many resources are being used with respect to what is available.

If this is out of scope for SO then I will remove.

Are there are any built in tools that do this or public scripts on GitHub that can achieve this functionality?

1

There are 1 best solutions below

1
Dom On

The command qstat -u "*" -nenv -j "*" outputs job details, including a line with job's usage: usage 1: wallclock=44:12:05:42, cpu=1:10:40:01, mem=9284973.79642 GBs, io=631.16018 GB, iow=65.130 s, ioops=22213570, vmem=284.719M, maxvmem=65.121G, rss=14.435M, ..., maxrss=61.611G, maxpss=68.641G

I am not aware of a public script that would parse it and cross reference the output of qhost to retrieve hosts resources.

I think I should be working on this over the weekend. :)