Andreas Haupt <[email protected]> writes: > Hi Mark, > > On Wed, 2011-02-23 at 09:20 +0000, Mark Dixon wrote: >> PS In case you're not already doing it: >> >> To check what the scheduler is thinking, "qconf -msconf" and add MONITOR=1 >> to the comma-separated list of values for "params". Look for "RESERVING" >> lines in file $SGE_ROOT/$SGE_CELL/common/schedule. > > Is the format of this file documented somewhere
I don't think so; I should do it if no-one gets there first. > 28009:1:RESERVING:1311972311:86700:H:pax66.ifh.de:tmpdir_size:8589934592.000000 > 28009:1:RESERVING:1311972311:86700:H:pax66.ifh.de:slots:8.000000 > 28009:1:RESERVING:1311972311:86700:H:pax66.ifh.de:h_vmem:25769803776.000000 > 28009:1:RESERVING:1311972311:86700:Q:[email protected]:slots:8.000000 > > it looks like this to me: > > <jobid>:<taskid>:<action>:<timestamp of reservation>:<reservation > duration>:<kind of resource>:<hostname>:<complex>:<amount of complex> > > Is the assumption correct? Roughly, at least. > Then it's interesting: Gridengine reserves > resources for the end of July ... That's a bit too late, I think ... Indeed. The time slots I've seen have been sensible (when the reservation has actually been there and not lost). It's clear at least something is broken in this area, with a range of symptoms, and we'll have to get it debugged somehow. _______________________________________________ users mailing list [email protected] https://gridengine.org/mailman/listinfo/users
