- 17 Mar, 2019 1 commit
-
-
Tim Wickberg authored
Not used at this point, so drop them from the sbatch/srun CLI. Leaving them in struct job_details for now, they may be removed at a future point if the rest of the checkpoint infrastructure is dropped.
-
- 16 Mar, 2019 6 commits
-
-
Tim Wickberg authored
Change mem_per_cpu / mem_per_gpu / pn_min_memory / pn_min_tmp_disk to uint64_t in slurm_opt_t, and use NO_VAL64 as the sentinel value throughout. Rework str_to_mbytes2() to return a uint64_t, and NO_VAL64 as an invalid value instead of -1.
-
Tim Wickberg authored
Not needed for now, may come back at a later time.
-
Tim Wickberg authored
Only used as part of an info() message hidden behind the --verbose option in salloc/sbatch/srun, which isn't worth keeping here as the rest of the slurm_opt_t structure is being cleaned up.
-
Danny Auble authored
# Conflicts: # NEWS # doc/man/man5/slurm.conf.5 # src/plugins/sched/backfill/backfill.c # src/slurmctld/job_scheduler.c
-
Morris Jette authored
This should help insure that jobs are not prevented from running at the same time on the same node due to memory limits. bug 6675
-
Danny Auble authored
-
- 15 Mar, 2019 9 commits
-
-
Tim Wickberg authored
Bug 6708.
-
Tim Wickberg authored
-
Dominik Bartkiewicz authored
when acct_policy_job_runnable_post_select() is call for job without job_resrcs (eg.: from _attempt_backfill() or when select_nodes() can't find free recource) but with req_node_bitmap. This patch will change reason of jobs like in test 21.39 from "JobState=PENDING Reason=QOSGrpNodeLimit" to "JobState=PENDING Reason=Resources" bug 6675
-
Tim Wickberg authored
-
Matt Ezell authored
Bug 6679
-
Tim Wickberg authored
-
Tim Wickberg authored
-
Tim Wickberg authored
-
Tim Wickberg authored
-
- 14 Mar, 2019 24 commits
-
-
Tim Wickberg authored
Promote two functions in opt.c to extern, and rename the entry point to the translation code to xlate_batch_script().
-
Tim Wickberg authored
-
Tim Wickberg authored
-
Tim Wickberg authored
Unused by launch/slurm.
-
Tim Wickberg authored
Only works with non-launch/slurm plugins.
-
Tim Wickberg authored
There is only launch/slurm currently.
-
Tim Wickberg authored
Hasn't been documented since at least version 2.6, and does nothing.
-
Tim Wickberg authored
Changed to the documented form before 17.11 was released.
-
Tim Wickberg authored
Return an xmalloc()'d string. Clean up string handling to use xstrfmtcat / xstrcat / snprintf instead of sprintf. Use that new function in a minimal new cpu_freq_set_env().
-
Morris Jette authored
Coverity CID 45165
-
Morris Jette authored
-
Morris Jette authored
-
Morris Jette authored
-
Moe Jette authored
No new functionality, just for clarity and consistency. Bug 6681
-
Felip Moll authored
Bug 6681
-
Morris Jette authored
-
Morris Jette authored
-
Morris Jette authored
Broken in commit 9118a41e
-
Morris Jette authored
Broken in commit 9118a41e
-
Tim Wickberg authored
-
Tim Wickberg authored
Can now only be used to launch steps in an existing job allocation; it can no longer be used by SlurmUser to create a new job allocation with a given jobid value. Bug 6691.
-
Tim Wickberg authored
Bug 6691.
-
Tim Wickberg authored
If running within salloc or a job step, both SLURM_JOB_ID and SLURM_JOBID would be set to the same value, and SLURM_JOB_ID would overwrite SLURM_JOBID as it is parsed second. The only time this would cause a difference in behavior is if SLURM_JOBID were set, and SLURM_JOB_ID were not. Which is not something that can be caused by Slurm itself.
-
Morris Jette authored
-