- 21 May, 2019 2 commits
-
-
Alejandro Sanchez authored
Node memory overallocation wouldn't be properly detected since we would just be interpreting the available memory as RealMemory - MemSpecLimit, ignoring other job's memory usage. Bug 5562.
-
Alejandro Sanchez authored
This compares a job memory request against each selected node available memory, interpreting the latter for now as RealMemory - MemSpecLimit. Bug 5562.
-
- 17 May, 2019 2 commits
-
-
Tim Wickberg authored
This is select/cons_res, not select/cons_tres.
-
Morris Jette authored
Previous select/cons_res logic would allocate one CPU per task on the node Bug 6981
-
- 16 May, 2019 1 commit
-
-
Marshall Garey authored
There was a syntax error in the mysql for inserting the event records into the event table caused by commit 3d61b6aa. The syntax error was a semicolon in the middle of the query, for example: insert into "voyager_event_table" (time_start, time_end, node_name, cluster_nodes, reason, reason_uid, state, tres) values ('1538669453', '1539298628', 'v1', '', 'cold-start', '1017', '0', '1=8,2=4000,5=8,1001=4,1002=1');, (<... another record>);, ... Bug 7025.
-
- 13 May, 2019 1 commit
-
-
Tim Wickberg authored
-
- 10 May, 2019 2 commits
-
-
Marshall Garey authored
Trying to archive too many records at once can result in archive files that are too big to read or even too big to be written. Only archive 50k records at a time, like we only purge 50k records at a time. Bug 6033.
-
Marshall Garey authored
The time period of the archive file currently depends on submit or start time and whether the purge period is in hours, days, or months. Previously, if the archive file name already exists, we would overwrite the old archive file with the assumption that these are duplicate records being archived after an archive load. However, that could result in lost records in a couple of ways: * If there were runaway jobs that were part of an old archive file's time period and are later fixed and then purged, the old file would be overwritten. * If jobs or steps are purged but there are still jobs or steps in that time period that are pending or running, the pending or running jobs and steps won't be purged. When they finish and are purged, the old file would be overwritten. Instead of overwriting the old file, we append a number to the file name to create a new file. This will also be important in an upcoming commit. Bug 6033.
-
- 06 May, 2019 1 commit
-
-
Felip Moll authored
When tres_usage_in_max field is empty it is recorded as '' in the database which leads find_tres_count_in_string() to return an INFINITE64. Seff treats INIFINITE64 as a valid value. This patch fixes this issue. Bug 6817
-
- 03 May, 2019 1 commit
-
-
Nate Rini authored
Bug 6880/6952.
-
- 02 May, 2019 2 commits
-
-
Broderick Gardner authored
On requeue, the origin cluster job record is copied to submit to sibling clusters. If the job was originally submitted to accept cluster default account, partition, etc, those fields are now filled in on the origin. Here we add flags to indicate that those fields need to be cleared on resubmission to siblings. Bug 6064
-
Broderick Gardner authored
This is a holdover from when the fed job_info list was added. The cluster lock has to be cleared from both the job_ptr and the job_info. Bug 6064
-
- 30 Apr, 2019 1 commit
-
-
Danny Auble authored
Blessed by Tim.
-
- 29 Apr, 2019 5 commits
-
-
Brian Christiansen authored
Bug 6513
-
Nate Rini authored
Bug 6895.
-
Brian Christiansen authored
Bug 6895
-
Brian Christiansen authored
Bug 6895
-
- 26 Apr, 2019 3 commits
-
-
Nate Rini authored
Otherwise, we could send communication packets bigger than max_allowed_packet. Bug 6832. Co-authored-by:
Tim Wickberg <tim@schedmd.com>
-
Alejandro Sanchez authored
Regression introduced in 8d643e79. Bug 6832.
-
- 24 Apr, 2019 3 commits
-
-
Moe Jette authored
when not the head job. Bug 6837 For more in depth explanation see comment 24
-
Albert Gil authored
When specifying -j and not -s then non-eligible jobs will be shown by sacct. Time windows can also be used with -S and -E. If --state is also used then non-eligible jobs won't be shown, because non-eligible are not actually PD. Bug 6873 # Conflicts: # NEWS
-
Ben Roberts authored
Bug 6805
-
- 23 Apr, 2019 2 commits
-
-
Danny Auble authored
Bug 6898
-
Albert Gil authored
Cancelled jobs before start have time_start=0 in the DB, but their time_end!=0 (the cancel time). Query fixed to handle properly these cases for -s PD. Bug 6894
-
- 22 Apr, 2019 1 commit
-
-
Ben Roberts authored
Brought the option in line with the "numtasks" option you use when not specifying --step. Preserved backwards compatbility for "numtask" for steps. Also upated docs and expect test. Bug 6861
-
- 18 Apr, 2019 4 commits
-
-
Dominik Bartkiewicz authored
Properly initialize structures throughout Slurm. Bug 6613
-
Danny Auble authored
Bug 6613
-
Dominik Bartkiewicz authored
Bug 6613
-
Tim Wickberg authored
Regression from aca37654 . Bug 6826. Co-authored-by:
Chad Vizino <chad@schedmd.com>
-
- 16 Apr, 2019 8 commits
-
-
Danny Auble authored
These are conditions that need to remain constant until something changes on the job to reevaluate. Bug 6625
-
Danny Auble authored
What was happening here is you had to not be >= operator to have the old limits removed. This makes it so it always happens. Bug 6625
-
Brian Christiansen authored
Bug 6625
-
Danny Auble authored
Before we went up the tree to the next assoc_ptr. As we validate an association on the id as well as the uid the assoc_ptr was eventually going to become invalid. Setting it to NULL here solves a bunch of issues with things later on. Bug 6625
-
Danny Auble authored
Bug 6625
-
Nathan Rini authored
Bug 6625.
-
Danny Auble authored
Don't abort when the job doesn't have an association that was removed before the job was able to make it to the database. Bug 6625
-
Brian Christiansen authored
Bug 6625
-
- 13 Apr, 2019 1 commit
-
-
Marshall Garey authored
The backfill scheduler keeps a local list of job pointers. Since the backfill scheduler yields locks, it's possible for pending jobs to be canceled and purged in these yield periods. The backfill scheduler then has pointers to now invalid memory, and dereferencing those pointers is undefined behavior and may result in a segfault. This commit prevents purging jobs while the backfill scheduler is running. Bug 6621
-