- 08 Dec, 2014 6 commits
-
-
David Bigagli authored
-
Brian Christiansen authored
Bug 1305
-
Morris Jette authored
Fix bug with GRES having multiple types that can cause slurmctld abort. This can be reproduced with select/cons_res and one Gres like this: Name=gpu Type=kepler File=/dev/tty0 A bad index was being used that caused an assert.
-
David Bigagli authored
-
Morris Jette authored
-
Artem Polyakov authored
Logic introdiced in version 14.03.10 to support requeueing of jobs with GRES allocated to currently running steps broke select/linear due to differernces in the plugin logic. The commit with the bad logic is 1209a664
-
- 05 Dec, 2014 5 commits
-
-
Brian Christiansen authored
Bug 1298
-
Brian Christiansen authored
-
Brian Christiansen authored
Bug 1301
-
Danny Auble authored
have no weight. This allows for association and QOS decay limits to work.
-
Danny Auble authored
Before without this option accounting would not be correct unless the job was allocating enough resources to fill up the socket. Making this the default makes it so the entire socket is allocated to job similar to the way CR_CORE works with allocating the entire core to the job even if they don't allocate the whole core.
-
- 04 Dec, 2014 6 commits
-
-
David Bigagli authored
draining in sinfo output.
-
Brian Christiansen authored
-
Brian Christiansen authored
Fix jobs from starting in overlapping reservations that won't finish before a "maint" reservation begins. Bug 1290
-
Morris Jette authored
Avoid huge malloc if GRES configured with "Type" and huge "Count".
-
Danny Auble authored
when the DBD is down.
-
Danny Auble authored
-
- 03 Dec, 2014 3 commits
-
-
Morris Jette authored
Log Cray MPI job calling exit() without mpi_fini(), but do not treat it as a fatal error. This partially reverts logic added in version 14.03.9. bug 1171
-
Brian Christiansen authored
Bug 1289
-
Danny Auble authored
could result in seg fault.
-
- 02 Dec, 2014 5 commits
-
-
Morris Jette authored
This only enables this configuration without providing the necessary infrastructure to support it. Add preempt_by_qos flag to select/cons_res In select/cons_res add extra row to scheduling array if preempt/qos is used. This is used only when a job is started via preempting another job in the same partition, but a lower priority QOS. Update preemption documentation
-
Danny Auble authored
better.
-
David Bigagli authored
-
Danny Auble authored
in BASIL was changed.
-
Brian Christiansen authored
-
- 28 Nov, 2014 1 commit
-
-
Dominik Bartkiewicz authored
In srun the default timeout in _wait_nodes_ready function is 3 sec. This patch make timeout adaptiv, for default max_delay=400 sec it make less rpc request and give beter 'user feelings'.
-
- 27 Nov, 2014 1 commit
-
-
Brian Christiansen authored
Bug #1276
-
- 26 Nov, 2014 5 commits
-
-
David Bigagli authored
in commit 1953fdecbb.
-
Morris Jette authored
-
Morris Jette authored
Job wait reason of "ReqNodeNotAvail" expanded to identify unavailable nodes (e.g. "ReqNodeNotAvail(Unavailable:tux[3-6])"). bug 1277
-
Morris Jette authored
-
David Bigagli authored
an error message and return 1.
-
- 24 Nov, 2014 2 commits
-
-
Morris Jette authored
Permit "SuspendTime=NONE" as slurm.conf value rather than only a numeric value to match "scontrol show config" output. This remove some special case logic in the test of "scontrol write config" and in the generation of the output file written.
-
Artem Polyakov authored
Double max string that Slurm can pack from 16MB to 32MB to support larger MPI2 configurations.
-
- 22 Nov, 2014 1 commit
-
-
Morris Jette authored
Added SchedulerParameters option of "bf_busy_nodes". When selecting resources for pending jobs to reserve for future execution (i.e. the job can not be started immediately), then preferentially select nodes that are in use. This will tend to leave currently idle resources available for backfilling longer running jobs, but may result in allocations having less than optimal network topology. This option is currently only supported by the select/cons_res plugin.
-
- 21 Nov, 2014 4 commits
-
-
David Bigagli authored
-
Danny Auble authored
-
Dominik Bartkiewicz authored
This can happen if the specified job ID is not found.
-
Morris Jette authored
Add advance reservation flag of "replace" that causes allocated resources to be replaced with idle resources. This maintains a pool of available resources that maintains a constant size (to the extent possible).
-
- 20 Nov, 2014 1 commit
-
-
Morris Jette authored
-