Fix karpenter configuration #45
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
…alf an hour. also adding long running gpu nodepool for long running training jobs
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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.
The issues of failing torch ddp runs on both CPU and GPU around the half hour mark incorrectly diagnosed as an argo context related issue was actually due to the karpenter nodepools being configured to remove all nodes after 30m no matter what.
This is now fixed, together with introducing short run vs long run GPU nodepools. the short run one should be used by default and currently supports all k8s tests w.r.t runtime.
Also removed the spot instances as they can get pulled as early as 30 minutes into the claim