Thanks for that info RE the code. FWIW, we have set up GPU resources the same in our cluster, and haven't run into that bug.
I wonder if deleting that execution host and adding it back again might work around your issue. Dan On Mon, May 14, 2018 at 12:52 PM, Joshua Baker-LePain <j...@salilab.org> wrote: > On Sun, 13 May 2018 at 8:49pm, Daniel Povey wrote > >> Can you show the full output from when you do `qstat -j <job-id>` for >> the job that's pending? > > > Unfortunately I had to change our setup so that GPU jobs would actually flow > through the queues -- we're no longer using a consumable gpu complex. Our > current setup, though, is far from perfect, which is why we're looking to > help get this fixed. > > In this <http://gridengine.org/pipermail/users/2018-April/010120.html> > message from the previous thread I mention that the 'qstat -j' output is > unremarkable. It details all the queues the job can't run in (all for > legitimate reasons). It's also notable that 'qalter -w p' always said > "verification: found possible assignment with 5 slots" when jobs got stuck > in this state. > > > -- > Joshua Baker-LePain > QB3 Shared Cluster Sysadmin > UCSF _______________________________________________ users mailing list users@gridengine.org https://gridengine.org/mailman/listinfo/users
