Avoid the brain split phenomenon in the symmetric network partition scenario #13221
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
OneSizeFitsQuorum
force-pushed
the
enhance_symmetric_network_partition_issue
branch
2 times, most recently
from
56b5ce4
to
c098f31
Compare
OneSizeFitsQuorum
force-pushed
the
enhance_symmetric_network_partition_issue
branch
from
c098f31
to
cb8e9fb
Compare
|
SteveYurongSu
approved these changes
Aug 19, 2024
CRZbulabula
approved these changes
Aug 19, 2024
Codecov ReportAttention: Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## master #13221 +/- ##
============================================
- Coverage 41.00% 40.84% -0.16%
Complexity 71 71
============================================
Files 3783 3825 +42
Lines 234996 236041 +1045
Branches 28140 28441 +301
============================================
+ Hits 96362 96422 +60
- Misses 138634 139619 +985 ☔ View full report in Codecov by Sentry. |
|
LGTM |
OneSizeFitsQuorum
deleted the
enhance_symmetric_network_partition_issue
branch
OneSizeFitsQuorum
added a commit
that referenced
this pull request
Aug 19, 2024
…cenario #13221 Signed-off-by: OneSizeFitQuorum <tanxinyu@apache.org>
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Outdated suggestions cannot be applied.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.
The current Ratis module's Leader may step down voluntarily without knowing who the new Leader is, which will not trigger the state machine's notifyLeaderChange callback. As a result, some modules that rely on this interface to determine whether the current node is no longer the Leader might delay resource release, potentially causing split-brain issues with multiple Leaders.
For example, in a 3-node ConfigNode setup, if a symmetric network partition fault is injected into the Leader node, the other two nodes will elect a new Leader. However, certain services (such as heartbeat, procedure, etc.) on the old Leader will not be cleared, leading to a split-brain scenario, which could cause some unexpected behavior.
After this PR, even if the new Leader is unknown, Ratis will still call the notifyNotReady function, thereby preventing split-brain issues from occurring.