- 02 Jul, 2018 3 commits
-
-
Morris Jette authored
-
Morris Jette authored
Implicitly set node count, task count, ntasks_per_* as appropriate
-
Morris Jette authored
Prior logic lacked a line-feed and ran lines together
-
- 29 Jun, 2018 5 commits
-
-
Morris Jette authored
-
Morris Jette authored
Previous logic was implicitly starting 2 tasks due to the combination of --gpus and --gpus-per-node
-
Morris Jette authored
-
Morris Jette authored
-
Morris Jette authored
-
- 27 Jun, 2018 1 commit
-
-
Morris Jette authored
Original logic was testing against number of available CPUs on a node, not the number actually being allocated. This can easily result in a valid job not being possible to run. Note that setting the max_cpus parameter requires the "scontrol update jobid=# numcpus=#-#" command be explicitly executed as the option is not available at job submit time, so it isn't something likely to happen on a production system. Also note this bug exists in the cons_res plugin and was just copied to cons_tres.
-
- 26 Jun, 2018 6 commits
-
-
Morris Jette authored
-
Morris Jette authored
-
Felip Moll authored
-
Felip Moll authored
When one asks for an inactive feature and also specifies the node with -w flag, the node will be rebooted despite it may contain running jobs. bug4821
-
Morris Jette authored
Return resource array structure rather than just CPU count
-
Tim Wickberg authored
and avoid race condition calling task before proctrack can introduce. Bug 5319
-
- 25 Jun, 2018 3 commits
-
-
Morris Jette authored
to work correctly. Bug 5155 Bug 4516
-
Morris Jette authored
delayed until the first job completes execution and it's burst buffer stage-out is completed. Bug 4675
-
Morris Jette authored
-
- 22 Jun, 2018 7 commits
-
-
Morris Jette authored
If MailProg is not configured and "/bin/mail" (the default) does not exist, but "/usr/bin/mail" does exist then use "/usr/bin/mail" as a default value.
-
Morris Jette authored
-
Morris Jette authored
-
Morris Jette authored
-
Morris Jette authored
Rather than eliminating cores in order across sockets, try to remove cores evenly across the sockets in order to preserve access to sockets which might have GPUs we want to use.
-
Dominik Bartkiewicz authored
Bug 5191
-
Dominik Bartkiewicz authored
Bug 5159.
-
- 21 Jun, 2018 3 commits
-
-
Morris Jette authored
This will clear unusable cores from the bitmap as needed.
-
Morris Jette authored
-
Morris Jette authored
-
- 20 Jun, 2018 12 commits
-
-
Morris Jette authored
Coverity CID 186635 and 186636 (same leak, two paths to the same place)
-
Morris Jette authored
Rather than calling select_g_job_test() multiple times as nodes are added for consideration by weight, put the node weight into a new sched_weight field in struct node_record and call select_g_job_test() once. Let it select the best resources, accumulating resources by increasing node weight.
-
Morris Jette authored
-
Tim Wickberg authored
MUNGE plugin is no longer packaged separate after the slurm.spec overhaul.
-
Morris Jette authored
bug 5185
-
Morris Jette authored
-
Morris Jette authored
-
Morris Jette authored
-
Morris Jette authored
-
Morris Jette authored
Enhancements to commit 35a13703 bug 5185
-
Brian Christiansen authored
Enhancements to commit 35a13703 bug 5185
-
Alejandro Sanchez authored
Previously the function was only testing against the first partition in the job_record. Now it detects if the job request is multi partition and if so then loops through all of them until the job will run in any or until the end of the list, returning the error code from the last one if the job won't run in any partition. Bug 5185
-