- 22 Jun, 2010 3 commits
-
-
-
Moe Jette authored
reservations.
-
-
- 21 Jun, 2010 8 commits
-
-
Moe Jette authored
-
Danny Auble authored
Fixed issue where if the slurmdbd was down past the MinJobAge and the start of the job hadn't been sent yet to throw it on the queue to be sent when the slurmdbd comes back up.
-
Moe Jette authored
-
-
Danny Auble authored
preemptive strike before 2.1 merge which will do this exact thing, but the code has changed locations so many conflicts will happen
-
Moe Jette authored
that of the slurmctld daemon: the same directory used for log files, SlurmdLogFile (if specified with an absolute pathname) otherwise the directory used to save state, SlurmdSpoolDir.
-
Moe Jette authored
-
-
- 18 Jun, 2010 14 commits
-
-
Danny Auble authored
-
Danny Auble authored
-
Moe Jette authored
affinity is enabled, keep track of the individual CPUs allocated to jobs rather than just the count of CPUs allocated (which could overcommit specific CPUs for running jobs).
-
Danny Auble authored
-
Phil Eckert authored
-
Phil Eckert authored
-
Moe Jette authored
-
-
Moe Jette authored
meaningful if StateSaveLocation=/dev/null
-
Moe Jette authored
-
Danny Auble authored
Fix for handling correctly jobs that resize, and also reporting correct stats on a job after it finishes.
-
Moe Jette authored
cleaner and tighter code
-
Moe Jette authored
-
Moe Jette authored
being freed by sched/backfill.
-
- 17 Jun, 2010 6 commits
-
-
-
Moe Jette authored
to provide detailed logging only if so configured and avoid a bunch of logging calls otherwise
-
Danny Auble authored
-
Danny Auble authored
-
Moe Jette authored
-
Moe Jette authored
-
- 16 Jun, 2010 2 commits
-
-
Moe Jette authored
for bit_rep_count (implcitly 1 for each node), also make the logic more clear
-
Don Lipari authored
i.e., Fixed a fair-share calculation bug in the priority/multifactor plugin
-
- 15 Jun, 2010 7 commits
-
-
Moe Jette authored
-
Moe Jette authored
to avoid having srun generate an error that the node isn't ready for use
-
Moe Jette authored
preempted jobs at the time slurmctld restarts that can result in over- subscribing resources.
-
Moe Jette authored
-
Moe Jette authored
-
Moe Jette authored
to eliminate the need for a cpu_rep_count argument and only require a single array of CPU counts per node. Modify src/slurmctld/gang.c to eliminate the extra array for simplicity, especially if we want to change CPU count per node on the fly.
-
Moe Jette authored
causes a job suspend/resume with substantial overhead.
-