Refactor deployments #196
Merged
Refactor deployments #196
Conversation
This file contains hidden or 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
MichaelRipa
reviewed
Nov 6, 2025
MichaelRipa
reviewed
Nov 6, 2025
MichaelRipa
reviewed
Nov 6, 2025
| dtype: str | torch.dtype = "bfloat16" | ||
|
|
||
|
|
||
| @ray.remote(num_cpus=2, num_gpus=0, max_restarts=-1) |
Member
There was a problem hiding this comment.
- Do we want to have
num_cpushardcoded? - Would it necessarily be bad to have a finite
max_restartsvalue (e.g 5)? What happens if we didn't have it set? It might make it easier to catch silent failures if the actor ends up eventually dying
MichaelRipa
reviewed
Nov 6, 2025
MichaelRipa
reviewed
Nov 6, 2025
MichaelRipa
reviewed
Nov 6, 2025
MichaelRipa
reviewed
Nov 6, 2025
| def purge(self): | ||
|
|
||
| for deployment in self.deployments.values(): | ||
| deployment.delete() |
Member
There was a problem hiding this comment.
I feel like deployment.delete() shouldn't handle exceptions itself. From a robustness standpoint, I think it's probably important for the node to know if an attempt to delete something failed.
Member
Author
There was a problem hiding this comment.
1.) What do you mean by "know" 2.) When would we need to know?
Member
There was a problem hiding this comment.
- Be aware whether an exception was raised in
deployment.delete() - When calling
deployment.delete()raises an exception
MichaelRipa
reviewed
Nov 6, 2025
MichaelRipa
reviewed
Nov 6, 2025
MichaelRipa
reviewed
Nov 6, 2025
MichaelRipa
reviewed
Nov 6, 2025
MichaelRipa
reviewed
Nov 6, 2025
MichaelRipa
reviewed
Nov 6, 2025
|
|
||
| del self.nodes[node] | ||
| node = self.nodes.pop(node_id) | ||
| node.purge() |
Member
There was a problem hiding this comment.
Similar to what I commented regarding deployment.delete(), I think any exceptions that happen in purge should propagate up here, and at the very least be logged as an error before proceeding
MichaelRipa
reviewed
Nov 6, 2025
MichaelRipa
reviewed
Nov 6, 2025
This was referenced Nov 10, 2025
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.
No description provided.