-
Notifications
You must be signed in to change notification settings - Fork 298
[live-migration] Fix source resume hang when cancelled before blackout #2871
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Merged
Harsh Rawat (rawahars)
merged 1 commit into
microsoft:main
from
rawahars:reuse-same-connection
Aug 12, 2026
+72
−46
Merged
Changes from all commits
Commits
File filter
Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
There are no files selected for viewing
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
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
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
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
Oops, something went wrong.
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Outdated suggestions cannot be applied.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.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Is there a potential race: connected is read here as a snapshot, but it is not synchronized with the recvLoopRoutine transition. If recvLoop() returns immediately before connected.Store(false) executes, Resume() can observe connected == true, set reuseLiveConn, and skip rebuilding the bridge. It can then clear migrating, after which recvLoopRoutine stores connected = false, observes migrating == false, and calls kill(err). This can leave the resumed session using a bridge that is subsequently terminated.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
In short, this workflow would be similar to how a pod/container is handled if the GCS connection was dropped randomly at any point today. Hence it should be safe, without introducing any backward incompatibility.
Long version:
So you are right that
IsBridgeConnectedis a point in time snapshot of the bridge state. Right now we are using it forResume APIwhich is a high level API from Controller. We can run into this API from 2 points-Destination Resume:rebuildBridgeisfalseand hence we would go into the loop. There is no prior bridge connection and therefore, we do that inside if.Source Rollback:blackout: VM was paused and hence listeners dropped and bridge would have collapsed duringTransferAPI causingconnectedto be false. Therefore, it would always return false here and lead us to go inside the if.blackout:Bridge is active at the time of Resume: connected is true and we skip the if.
As soon as the
c.guest.IsBridgeConnected()is read inside Resume, the VM terminates or connection drops for some reason.This is the race scenario but it's safe and follows the existing workflow. If this happens then we would mark the bridge as not migrating and finish the
finalize. Since the GCS connection is not present, the bridge collapses onmigrating == falseand all the processes/containers signal their EXIT.containerdgets the notification and trigger delete. The caller sees the container asEXITEDand pod asNOTREADY. Caller can then shut the shim.This would be similar to the scenario where no migration is going on and the GCS connection dropped mid-way which causes the bridge to collapse and all processes to trigger their
EXIT.