slfan1989 opened a new pull request, #1356: URL: https://github.com/apache/ratis/pull/1356
## What changes were proposed in this pull request? This PR fixes a bug in `NettyRpcProxy.Connection.offer()` where `CompletableFuture` objects were added to the replies queue before sending requests via `writeAndFlush()`. When `writeAndFlush()` failed (either synchronously or asynchronously), the futures were not removed from the queue, causing: - **Memory leak**: Unreplied futures accumulate in the queue, holding references to large proto messages - **Reply mismatch**: Subsequent responses are matched to wrong futures, causing Raft protocol errors and potential data corruption **Changes made:** - Added try-catch in `NettyRpcProxy.Connection.offer()` to remove the future from queue when `writeAndFlush()` throws `AlreadyClosedException` - Added a `ChannelFuture` listener to handle asynchronous write failures and clean up the queue - Added unit test `TestNettyRpcProxy.testOfferRollbackOnAlreadyClosed()` to verify the fix ## What is the link to the Apache JIRA JIRA: RATIS-2415. Fix queue corruption in NettyRpcProxy when request sending fails. ## How was this patch tested? Added `TestNettyRpcProxy.testOfferRollbackOnAlreadyClosed()` -- 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]
