wip: kv: evict leaseholder on RPC error #23938
Closed
+4
−1
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
Which is what previously happened, thanks to the call to t.Fatal which in turn called runtime.Goexit. We need to call `done` here so that the caller has their WaitGroup signaled. Release note: None
|
This change is |
|
Typed faster than my eyes could follow. |
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Outdated suggestions cannot be applied.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.
This addresses a situation in which we would not evict a stale
leaseholder for a long time. Consider the replica [s1,s2,s3] and
s1 is down but is the cached leaseholder, while s2 is the actual lease
holder. The RPC layer will try s1, get an RPC error, try s2 and succeed.
Since there is no NotLeaseHolderError involved, the cache would not get
updated, and so every request pays the overhead of trying s1 first.
WIP because needs testing.
Touches #23601.
Release note (bug fix): Improve request routing during node outages.