- 24 Oct, 2014 11 commits
-
-
Morris Jette authored
-
Morris Jette authored
If the slurm.conf file includes other files and those files do not include an absolute path name, they will be taken from the same directory as slurm.conf. Previously it used the search path.
-
Morris Jette authored
-
Morris Jette authored
Conflicts: testsuite/expect/test2.18
-
Morris Jette authored
Use default_partition by default for the test
-
Morris Jette authored
Returned test passed, when it should have failed because of bad variable name.
-
David Gloe authored
This matches the value in version 14.03
-
Morris Jette authored
-
Morris Jette authored
-
David Singleton authored
We've seen slurmctld crashes due to negative job array indices.
-
Morris Jette authored
add field to slurmctld job structure, save/restore field, Add support to view/set in scontrol and sview Add support to view in squeue Validate job update string lengths
-
- 23 Oct, 2014 13 commits
-
-
Morris Jette authored
Add "--bb" (burst buffer specification) option to salloc, sbatch, and srun. Information reaches slurmctld, but is not processed more (yet).
-
Morris Jette authored
-
Morris Jette authored
-
Morris Jette authored
-
Morris Jette authored
The previous patch should work in most cases, but this should work more reliably and the comment is more clear bug 1196
-
Danny Auble authored
-
Danny Auble authored
it was looking at 08 as an octal number and not printing it correctly. This patch forces it to be looked at as a base 10 int. I did the same thing for the SLURM_API_VERSION just to be safe since it used the same kind of logic.
-
David Bigagli authored
This reverts commit 7e65f924.
-
David Bigagli authored
-
Morris Jette authored
-
Morris Jette authored
Also don't fail a job that could not be requeued because it was already pending. This was not actually observed, but appeared to be a possible race condition discovered in working on bug 1196
-
Morris Jette authored
-
Morris Jette authored
BGQ: Fix race condition when job fails due to hardware failure and is requeued. Previous code could result in slurmctld abort with NULL pointer. bug 1096
-
- 22 Oct, 2014 5 commits
-
-
David Bigagli authored
-
Gennaro Oliva authored
-
Morris Jette authored
Conflicts: src/slurmctld/job_mgr.c
-
David Gloe authored
The slurm_sbcast_lookup() API was recently changed to take a step ID. This changes the Perl API to match that.
-
Morris Jette authored
If a job is being tested to determine if it can ever run, then do not set the reason based upon account/QOS limits since that reason will likely be wrong anyway. bug 1186
-
- 21 Oct, 2014 9 commits
-
-
David Bigagli authored
-
Morris Jette authored
-
Don Lipari authored
Let operators (in addition to user root and SlurmUser) see job script for other user's jobs. bug 1176
-
Nicolas Joly authored
bug 1195
-
Nicolas Joly authored
-
Morris Jette authored
-
Morris Jette authored
-
Morris Jette authored
Fix bug that prevented preservation of a job's GRES bitmap on slurmctld restart or reconfigure (bug was introduced in 14.03.5 "Clear record of a job's gres when requeued" and only applies when GRES mapped to specific files). bug 1192
-
David Bigagli authored
-
- 20 Oct, 2014 2 commits
-
-
Danny Auble authored
-
Danny Auble authored
-