andyhuangdev opened a new pull request, #11903: URL: https://github.com/apache/ozone/pull/11903
## What changes were proposed in this pull request? Container reconciliation now always verifies checksums when reading chunks from a peer, regardless of the datanode’s ozone.client.verify.checksum setting. Previously, disabling this setting allowed reconciliation to copy corrupt peer data and mark it healthy in the local Merkle tree. A subsequent healthy peer could not repair that chunk during the same task, and the final scan could mark the replica UNHEALTHY. The change enables checksum verification on the configuration passed to the reconciliation stream, following the existing EC reconstruction approach. Corrupt chunks are rejected before being written, allowing reconciliation to repair the replica from another peer. ## What is the link to the Apache JIRA [HDDS-16805](https://issues.apache.org/jira/browse/HDDS-16805) ## How was this patch tested? Added a regression test with client checksum verification disabled, a corrupt peer with stale scan results, and a healthy peer. The test failed before the fix and now verifies that reconciliation restores healthy data and leaves the replica CLOSED. All 18 tests in TestContainerReconciliationWithMockDatanodes pass. Repository checkstyle and git diff --check also pass. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
