[native] Cancel pending async callback before resetting client and evb #24308
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
yuandagits
force-pushed
the
periodicCrash
branch
5 times, most recently
from
a6854ee to
825a127
Compare
yuandagits
force-pushed
the
periodicCrash
branch
from
825a127 to
97339a6
Compare
xiaoxmeng
approved these changes
Dec 30, 2024
xiaoxmeng
changed the title
|
There appears to be an unrelated test failure |
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Outdated suggestions cannot be applied.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.
Description
There is a rare race at shutdown time when annoucer_->stop() is called which will reset fields like client_ and the evb_. However, this is race-y because the libevent timer on the evb can go-off, and we set stop_ = true even though sendRequest() has progressed beyond that point. Essentially, it suffers from time-of-check to time-of-use race.
To fix, we can cancel all pending events on the evb.
Note, you may wonder why the comment in
sendRequestdoesn't work.This is because client_ depends on evb_, so client must be destructed before evb_ due to dependency issue, otherwise there will be SIGSEGV in client_'s d-tor. This destruction order is actually not desirable because we want the evb_ to finish all of the clean up before client is destructed, but we cannot because of the order issue.
This is why we must explicitly call cancel on the evb_'s timer itself.
Motivation and Context
Impact
Test Plan
Deploy to test cluster and restarted 4 times without any issues.
Contributor checklist
Release Notes
Please follow release notes guidelines and fill in the release notes below.
If release note is NOT required, use: