Numerous improvements to Slurm reboot handling and planned node state #115
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This takes work from #84 and expands on it a bit. Now something like this:
Will reboot a node and when it comes up it won't get marked online until NHC deems the node healthy. This only works by NOT defining the
reason
parameter to this scontrol command and also ensuring the nextstate is DOWN. This is especially a useful feature when doing rolling reboots using Slurm to avoid nodes coming online too quickly before some services like GPFS have had a chance to start.There is also an addition in here to better support unexpected reboots of nodes when using ReturnToService=1.
Also treat planned nodes like nodes that are essentially idle as planned nodes are idle nodes held back for future jobs.
These patches have been deployed to Ohio Supercomputer Center for quite a while now.