Skip to content

[To dev/1.3] Fix ConfigNode WAL and node registration (#17609) - #18934

Open
Caideyipi wants to merge 1 commit into
apache:dev/1.3from
Caideyipi:codex/cp-08380be-dev-1.3
Open

Caideyipi wants to merge 1 commit into
apache:dev/1.3from
Caideyipi:codex/cp-08380be-dev-1.3

Conversation

@Caideyipi

Copy link
Copy Markdown
Collaborator

Description

ConfigNode could apply SimpleConsensus plans before their WAL records were durable and report successful node registration after failed consensus writes. This backport persists WAL records before execution and propagates registration failures, rolling back incomplete DataNode/AINode registration.

Backport #17609 (08380be95ae82bffc9e1aa8adf7cc1a92eb027ee) to dev/1.3.

  • Roll back failed WAL appends, force operations, and plan execution; fix log rotation, recovery indexes, and writer shutdown.
  • Keep ConfigNode registration pending for retry after persistence failure and notify success after both registration plans persist.
  • Preserve successful snapshots when pipe listener processing fails, and fix partition counting and maintenance task snapshots.

Adapted to the 1.3 consensus request, snapshot loading, and runtime configuration APIs. Includes the EN/ZH message constants used by this fix with locale source selection in the ConfigNode module.

Validation

All 44 reactor modules passed the English build and Chinese test compilation. The four affected test classes passed all 19 regression tests.

mvn spotless:apply -pl iotdb-core/confignode
mvn -nsu test '-Dtest=ConfigRegionStateMachineTest,NodeManagerRegistrationTest,NodeInfoTest,PartitionInfoTest' '-Dsurefire.failIfNoSpecifiedTests=false'
mvn -nsu test-compile -DskipTests -P with-zh-locale

Also verified EN/ZH key and placeholder parity for all 21 message constants and git diff --check origin/dev/1.3.

This PR has:

  • been self-reviewed.
  • added regression tests for WAL failures, registration failures, WAL replay, and partition counting.

* Fixed multiple config bugs

* fix

* Update ConfigRegionStateMachine.java

* Address ConfigNode review feedback

* Fix compile

* Compile fix

* Fix NodeManager Sonar issues

* Fix ConfigNode WAL recovery and registration review feedback
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant