Hi, Am 10.04.2013 um 11:21 schrieb Arnau Bria:
> I'd like to know what temporally directories where not removed after > job finished in a node . > > I've created an scripts that checks $TMPDIR and compares its contents > with the job_scripts content (I assumed that all the jobs have a > job_script associatd), but now I've found that my assumption is not > true in all cases. What do you want to compare there in detail - you parse the job script (by it's JOB_ID) and scan for mentioned files? Inside any called binary or "additional shell script" files can be created too. In addition: SGE will store some in information in $TMPDIR too (in case of a parallel program). > One node is running a 20 slots job and it does not have a job_script: > > sgeadmin 17707 0.0 0.0 65304 6032 ? S Apr02 0:00 \_ > sge_shepherd-295294 -bg > > # ls -lsa /var/spool/gridengine/node-ib0002bi/job_scripts/ > total 8 > 4 drwxr-xr-x 2 sgeadmin sgeadmin 4096 Mar 26 21:57 . > 4 drwxr-xr-x 5 sgeadmin sgeadmin 4096 Mar 24 03:38 .. > > but it has an entry under active_jobs : > > /var/spool/gridengine/node-ib0002bi/active_jobs/295294.1/ > > So, few questions here: > > 1.-) not all the jobs are supposed to have a job_script? Correct. If you submit with "-b y" you can call a binary directly without any jobscript. > 2.-) if 1), may I trust active_jobs instead of jobs_scripts? Yes. > Extra question: > > how do other admins control the disk space used per job under $TMPDIR ? On the exechost? I don't do it at all on a per job basis. In case your users fight for the disk space you can implement a consumable for the disk space in combination with a load sensor: http://gridengine.org/pipermail/users/2012-February/002914.html (there are some other points in the thread too, like mounting a limited loop device on $TMPDIR) > any way for limiting the amount of space per job (SGE or OS level)? Not that I'm aware of in Linux. -- Reuti > TIA, > Arnau > _______________________________________________ > users mailing list > [email protected] > https://gridengine.org/mailman/listinfo/users _______________________________________________ users mailing list [email protected] https://gridengine.org/mailman/listinfo/users
