Conversation
This will outright reject jobs that requests more nodes or more than MaxMemPerCPU, MaxMemPerNode, MinNodes, MaxNodes, MaxTime, AllocNodes, AllowAccounts, AllowGroups, AllowQOS, and QOS usage than is available in any of the requested partitions. If submitting to multiple partitions and one partitions fits, it will be used (contrary to EnforcePartLimits=ALL).
![gemini-code-assist[bot]](https://avatars.githubusercontent.com/in/956858?s=60&v=4){kind=small}
There was a problem hiding this comment.
Code Review
This pull request introduces EnforcePartLimits: ANY as a new default for the Slurm configuration. This is a sensible change that improves user experience by providing faster feedback on job submissions. The implementation is correct. I've added a specific comment regarding test coverage. Additionally, since this changes a default behavior, please consider documenting this in the project's CHANGELOG or release notes to inform users.
| SlurmctldSyslogDebug: info | ||
| SlurmdSyslogDebug: info | ||
| PropagateResourceLimitsExcept: MEMLOCK | ||
| EnforcePartLimits: ANY |
There was a problem hiding this comment.
While this is a sensible new default, changes to default behavior should be accompanied by tests to prevent future regressions. Please consider adding a test that verifies this configuration is correctly applied, for instance by checking the output of scontrol show config in a molecule test.
1 participant
This will outright reject jobs that requests more nodes or more than MaxMemPerCPU, MaxMemPerNode, MinNodes, MaxNodes, MaxTime, AllocNodes, AllowAccounts, AllowGroups, AllowQOS, and QOS usage than is available in any of the requested partitions.
If submitting to multiple partitions and one partitions fits, it will be used (contrary to EnforcePartLimits=ALL).