the goal is to make a more vfx related viewer to the gridengine that doesn't rely on parsing qstat or qacct (which i have done in the past) and doesn't rely on wrapping qsub to create initial entries in my database.
i'll also be storing a lot of information not related to sge in the database for the task. the type if application it is. comment from the artists, which files it's using, which files it's creating,etc etc. accounting only gives me data after the fact. I would like to know while the machine accepts the task so i can visualise the progress to my artists. On Fri, Aug 19, 2016 at 3:53 PM, Reuti <re...@staff.uni-marburg.de> wrote: > Hi, > > > Am 19.08.2016 um 15:58 schrieb Lars van der Bijl <com48...@gmail.com>: > > > > hey all, > > > > I'm trying to parse the reporting file and parse this data into a > mongodb. > > What is the goal of this effort? There is also the accounting file which > contains the execd the job was executed on (`man accounting`). > > -- Reuti > > > > in the reporting with the joblog on we get message that look like: > > > > 1471613763:job_log:1471613763:sent:203471:1140:NONE:t: > master:qmaster:0:924:1471613584:jobname:lars:users:job:defaultdepartment:sge:sent > to execd > > > > 1471613763:job_log:1471613763:delivered:203471:1140:NONE:r: > master:qmaster:0:974:1471613458:jobname:lars:users:job:defaultdepartment:sge:job > received by execd > > > > looking at the code of Son of Grid Engine I see a reference to > qualified_hostname being given to reporting_create_job_log. but this is > the qmaster. > > > > would it be possible to add the execd it was sent to? > > > > or. is there a better way to get a detailed view of the cluster I can > parse and update my own database with? > > _______________________________________________ > > users mailing list > > users@gridengine.org > > https://gridengine.org/mailman/listinfo/users > >
_______________________________________________ users mailing list users@gridengine.org https://gridengine.org/mailman/listinfo/users
