- 11 Aug, 2016 3 commits
-
-
Morris Jette authored
-
Morris Jette authored
Don't about step launch if job reaches expected end time while node is configuring/booting (NOTE: The job end time will be adjusted after node becomes ready for use). bug 2985
-
Morris Jette authored
The sleep starts at 0.1 second and doubles each time to a maximum value. Previous maximum was 2 seconds, increased to 5 seconds to account for very large KNL boot times (tens of minutes) and to limit overhead on slurmctld daemon.
-
- 10 Aug, 2016 10 commits
-
-
Morris Jette authored
If the KNL node modes reported by cnselect and capmc differ, log that using "info" rather than "debug".
-
Morris Jette authored
Optionally lock slurmstepd in memory for performance reasons and to avoid possible SIGBUS if the daemon is paged out at the time of a Slurm upgrade (changing plugins). Controlled via new LaunchParameters options of slurmstepd_memlock and slurmstepd_memlock_all. bug 2334
-
Danny Auble authored
frequency other than AcctGatherNodeFreq.
-
Danny Auble authored
Filesystem instead of Lustre.
-
Iakovos Panourgias authored
Network dataset.
-
Morris Jette authored
-
Morris Jette authored
Locking slurmstepd in memory can result in exhausting real memory in some cases, resulting in failure of the slurmstepd process. This reverts commit 03cf4a5d, but the logic will be returned using a configuration parameter in Slurm version 17.02. bug 2334
-
Morris Jette authored
-
Morris Jette authored
This should improve performance and prevent failure if a local group ID lookup fails. bug 2928
-
Alejandro Sanchez authored
Minor correction to commit 30baec8d
-
- 09 Aug, 2016 27 commits
-
-
Tim Wickberg authored
-
Morris Jette authored
Prevent slurmd abort if hwloc library fails to populate the "children" arrays (observed with hwloc version "dev-333-g85ea6e4").
-
Morris Jette authored
-
Morris Jette authored
-
Morris Jette authored
-
Danny Auble authored
-
Danny Auble authored
4 billion jobs in your database you will want to update your slurmctld at the same time as your slurmdbd to prevent roll over of this variable as it is 32 bit previous versions of Slurm.
-
Danny Auble authored
mysql_insert_id was always returning a my_ulonglong so we are just making it more compliant. This is needed primarily for a following commit that will change the job db_index to be be 64bit.
-
Tim Wickberg authored
Bug 2955.
-
Morris Jette authored
-
Morris Jette authored
Document the partition limits considered in the EnforcePartLimits logic. bug 2920
-
Morris Jette authored
-
Morris Jette authored
Correction to commit b45a55bc but 2920
-
Morris Jette authored
-
Morris Jette authored
Make EnforcePartLimit support logic work with any ordering of partitions in job submit request. Developed jointly with Alejandro Sanchez <alex@schedmd.com> bug 2920
-
Tim Wickberg authored
Copy + paste error, CID 44786.
-
Tim Wickberg authored
-
Dominik Bartkiewicz authored
Calculation switched the node count in place of the cpu count, which results in incorrect estimates. CID 44784.
-
Dominik Bartkiewicz authored
CID 44787.
-
Dominik Bartkiewicz authored
Mistake was only visible with DEBUG_TIME, and under a debug3 message. CID 44785.
-
Morris Jette authored
-
Tim Wickberg authored
Bug 2950. Also identified as CID 56684 (copy+paste error).
-
Morris Jette authored
-
Dominik Bartkiewicz authored
CID 45023 and 45024.
-
Tim Wickberg authored
This now provides a consistent approach to when all partition priorities become zero. Before 'scontrol update' could leave norm_prio on the last partition to become zero as one.
-
Tim Wickberg authored
-
Tim Wickberg authored
Convert _calc_part_tres into a ListForF signature. Also removes a double xfree of part_ptr->tres_cnt.
-