HBASE-30353 Prevent split parent from being re-opened after master failover - #267
Draft
tanyastickles wants to merge 6 commits into
Draft
tanyastickles wants to merge 6 commits into
tanyastickles wants to merge 6 commits into
Conversation
…ilover After a split, MetaTableAccessor.splitRegion writes split=true/offline=true into info:regioninfo but never updates info:state from CLOSED (left by the pre-split unassign). On failover, loadMeta rebuilds the RegionStateNode with state=CLOSED, so preTransitCheck accepted the parent for assign because CLOSED is in STATES_EXPECTED_ON_ASSIGN. The parent could then be opened, archiving store files the daughters still reference via HFileLinks. Fix: add RegionStateNode.checkNotRetired(), which throws DoNotRetryRegionException when isSplit() is true. Call it from preTransitCheck before any external assign can proceed, covering both the live state (state=SPLIT) and the post-failover state (regionInfo.isSplit()=true, state=CLOSED). Adds TestSplitParentAssignment to reproduce the exact post-failover precondition by deleting the existing RSN and re-creating it from a split=true RegionInfo with state=CLOSED, then asserting that assign() throws DoNotRetryRegionException. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
…h; add unit tests - Add checkNotRetired() before the override branch in createAssignProcedure so HBCK2 and other override=true callers cannot assign a split parent (previously only preTransitCheck covered the override=false path) - Port two mock-based unit tests from the alternate branch into TestAssignmentManager: CLOSED-after-failover and SPLIT-state cases, each asserting both assign() throws and createOneAssignProcedure returns null Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Follow WET guidance: inline the isSplit() check directly in AssignmentManager at both call sites instead of abstracting it into RegionStateNode.checkNotRetired(). Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
…ass name) Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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
Upstream Jira: HBASE-30353
Internal Issue: https://github.com/HubSpotEngineering/HBasePlanning/issues/2627
We ran into a bug where if the HMaster fails over after a region is split, a split parent can come online again. After a region split, a split parent should never be opened.
Step by step:
state=SPLIT. In the metadata, it writessplit=true, offline=true, state=CLOSED.state=CLOSED. The parent region ends up looking like an ordinary stuck in closing region instead of a split parent.BRAVE
Backwards Compatibility
Rollout and Rollback Plan
Automated Testing
Verification
Expect Dependencies to Fail
REVIEWERS: Please review both the code changes and the answers above, and validate that they match the expectations for BRAVE
🤖 Generated with Claude Code