Repository navigation
atelet: a detach whose upload attempt times out is Unavailable - #3532
Open
Davanum Srinivas (dims) wants to merge 1 commit into
Open
Davanum Srinivas (dims) wants to merge 1 commit into
Davanum Srinivas (dims) wants to merge 1 commit into
Conversation
Davanum Srinivas (dims)
force-pushed
the
detach-upload-timeout
branch
from
November 7, 2026 03:26
4216d1c to
106c2ea
Compare
Michelle Au (msau42)
left a comment
Michelle Au (msau42)
left a comment
Collaborator
There was a problem hiding this comment.
lgtm, just one test comment
Davanum Srinivas (dims)
force-pushed
the
detach-upload-timeout
branch
from
November 13, 2026 03:07
106c2ea to
baa4ec9
Compare
At 6x and 8x on microVM every resume of an actor whose node is full detaches it, and dozens of snapshot uploads run against the object store at once. An attempt that runs out of its own time keeps the local copy and the tracker retries it, but the plain error reached the API as Internal and the router answered 500, which the client cannot retry. Report it as Unavailable: the API already maps that to a retryable failure and the router keeps the request parked. Signed-off-by: Davanum Srinivas <davanum@gmail.com>
Davanum Srinivas (dims)
force-pushed
the
detach-upload-timeout
branch
from
November 13, 2026 03:09
baa4ec9 to
02f24b8
Compare
Collaborator
Author
|
Michelle Au (@msau42) fixed. PTAL |
This branch has not been deployed
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Under load many actors on a full node are resumed at once. Each resume detaches the actor first, which uploads its snapshot, so dozens of uploads run against the object store at the same time. An attempt can run out of its own upload timeout.
The tracker keeps the local snapshot and retries the upload on the next attempt. But the error it returned had no gRPC code, so the detach failed as Internal:
The router answered 500, which the client cannot retry. One case of #646.
Return Unavailable when the attempt timed out while the caller's context is still live.
detachActorErrormaps Unavailable to a retryable failure, so the router keeps the request parked and retries the resume.Written with Claude Code: it found the path in api-server logs and wrote the change and the test.