kvserver: load committed entries after mu.Unlock() #143689
Open
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
![blathers[crl]](https://avatars.githubusercontent.com/in/59700?s=40&u=95b789b1d594ac6b74f377aa7555d8d98d63240d&v=4){kind=avatar}
|
It looks like your PR touches production code but doesn't add or edit any test code. Did you consider adding tests to your PR? 🦉 Hoot! I am a Blathers, a bot for CockroachDB. My owner is dev-inf. |
|
This change is |
|
@tbg Would you be interested in measuring the impact of this on #140235? Or give me a hint how to do it. A conjecture is that it should have some impact because all these entries are not in |
pav-kv
force-pushed
the
load-committed-under-raft-mu
branch
from
64759f3 to
79a69b5
Compare
pav-kv
commented
Mar 31, 2025
pav-kv
force-pushed
the
load-committed-under-raft-mu
branch
from
79a69b5 to
60dc887
Compare
pav-kv
commented
Mar 31, 2025
This commit makes handleRaftReadyRaftMuLocked load the Ready.Committed entries after releasing Replica.mu, while still holding Replica.raftMu. This ensures that we don't have this IO under Replica.mu. This also fixes a bug. Previously, detachRaftEntriesMonitorRaftMuLocked would be called too soon, after Ready has been generated. Instead, it should span the entire Ready handling scope, so that includes the scope of applying the entries to the state machine. Epic: none Release note: none
pav-kv
force-pushed
the
load-committed-under-raft-mu
branch
from
60dc887 to
a630095
Compare
pav-kv
commented
Mar 31, 2025
|
|
||
| logRaftReady(ctx, rd) | ||
| outboundMsgs, msgStorageAppend = splitLocalStorageMsgs(rd.Messages) | ||
| ready = raftGroup.Ready() |
There was a problem hiding this comment.
There are a bunch of redundant checks between HasReady() and Ready() calls. We could further squash this in the future as:
ready, hasReady = raftGroup.Ready()
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 commit makes
handleRaftReadyRaftMuLockedload theReady.Committedentries after releasingReplica.mu, while still holdingReplica.raftMu. This ensures that we don't have this IO underReplica.mu. Also, this storage interaction is placed afterr.sendRaftMessages, which reduces messaging latency.This also fixes two bugs.
detachRaftEntriesMonitorRaftMuLockedwould be called too soon, afterReadyhas been generated. Instead, it should span the entireReadyhandling scope, so that includes the scope of applying the entries to the state machine.raftMu.bytesAccountwas unsafe to refer under onlyReplica.mu.Part of #140235
Related to #143652, #143615
Obsoletes #125842