- Jan 03, 2017
-
-
Dominik Bartkiewicz authored
PluginDir is allowed to be a PATH-style list of directories; remove incorrect test of the variable as if it were a single directory and comment that the check for that is elsewhere. Bug 3361.
-
Tim Wickberg authored
-
Alejandro Sanchez authored
-
Morris Jette authored
Expect is unable to work with more than 32-bit numbers, so it can't properly test task affinity on KNL with 272 threads
-
Tim Wickberg authored
-
Alejandro Sanchez authored
- 2 x Integer handling issues (CONSTANT_EXPRESSION_RESULT)
-
Tim Wickberg authored
Warning flag -Wempty-body in clang-3.7 now errors on this, split macro definition up over multiple lines to silence.
-
- Dec 31, 2016
-
-
Morris Jette authored
Revision to 06fbdd75 Previous logic was causing tests 1.87 and 15.19 to fail sometimes due to a race condition by adding a "Job allocation revoked" message to salloc output even if no allocation occured. bug 3351
-
- Dec 29, 2016
-
-
Roland Fehrenbacher authored
-
Morris Jette authored
Previous logic was failing if no specialized cores configured on node and job wanted to configure some bug 3349
-
Dominik Bartkiewicz authored
Null terminate before strchr().
-
Morris Jette authored
-
Morris Jette authored
This is a new message when "PrologFlags=contain" or "PrologFlags=alloc" is configured. bug 3351
-
- Dec 28, 2016
-
-
Alejandro Sanchez authored
Cancel interactive job if Prolog failure with "PrologFlags=contain" configured. bug 3351
-
Morris Jette authored
Minor fix to commit 1e6b7522 in the case where user specification of specialized cores is disabled. bug 3349
-
Matthieu Hautreux authored
If a job specifies --core-spec and the node has identified specialized cores using the CpuSpecList option which are not the default cores (e.g. the first cores rather than the last cores on each socket), the the user's cores will be selected from those identified in the configuration. bug 3349
-
- Dec 27, 2016
-
-
Morris Jette authored
Was printing int64 using %ld, which generated a warning on some systems
-
- Dec 21, 2016
-
-
Morris Jette authored
Do not allocate specialized cores to jobs using the --exclusive option. bug 3349
-
- Dec 19, 2016
-
-
Morris Jette authored
This commit should be part of b5f621d0 when the man page was removed
-
Morris Jette authored
Fix memory and file descriptor leaks in slurmd daemon's sbcast logic.
-
- Dec 16, 2016
-
-
Danny Auble authored
The part_ptr is sent into the function, there is no reason to look it up again. Coverity reported this.
-
Tim Wickberg authored
TravisCI uses 2-cpu nodes, so this causes the test to fail as a suitable node cannot be found.
-
Tim Wickberg authored
-
Tim Wickberg authored
Plugins are deprecated and will be removed in 17.02.
-
- Dec 15, 2016
-
-
Danny Auble authored
version is lower than the min version, set it to the min. Bug 3050
-
Alejandro Sanchez authored
-
Morris Jette authored
Address new coverity error report
-
Morris Jette authored
sched/backfill - Fix logic to reserve resources for jobs that require a node reboot (i.e. to change KNL mode) in order to start. bug 3346
-
Danny Auble authored
go into JobAdminHeld. Bug 3201
-
- Dec 14, 2016
-
-
Morris Jette authored
Fix for possible infinite loop in select/cons_res plugin when trying to satisfy a job's ntasks_per_core or socket specification. bug 3329
-
Tim Wickberg authored
Bug 2992.
-
Morris Jette authored
Modify regression test1.89 to avoid leaving vestigial job. Also reduce logging to reduce likelyhood of Expect buffer overflow. bug 3273
-
- Dec 13, 2016
-
-
Morris Jette authored
-
Tim Wickberg authored
Reverts most of commit 84023f27. Searching the PATH in slurmd can fail due to root_squash'd NFS filesystems, leading to the "wrong" program being launched. If you'd like the performance benefit from avoiding this lookup during each separate task launch, set SLURM_TEST_EXEC=1 instead which will perform the lookup once within srun, which then ensures the lookup happens under the users own environment and not that of the slurmd. Bug 2992.
-
Tim Wickberg authored
-
- Dec 12, 2016
-
-
Alejandro Sanchez authored
-
Alejandro Sanchez authored
-
Morris Jette authored
If running burst_buffer/cray plugin on emulated Cray (NOT a real Cray system with Cray APIs) and the burst buffer state information includes information about buffers associated with jobs which are either completed or no longer exist (i.e. the burst buffer state information is old), then ignore that old state information. This might possibly be related to bug 3295 if their configuration is bad.
-
- Dec 09, 2016
-
-
Danny Auble authored
level.
-
- Dec 08, 2016
-
-
Danny Auble authored
-