ramanathan1504 commented on PR #4319: URL: https://github.com/apache/logging-log4j2/pull/4319#issuecomment-5909941769
> Agreed: a failure in connectAndStart() must keep the pending batch because those events have not reached the database. I propose capping it at bufferSize by retrying a full batch before accepting another event; if that retry still fails, the write throws before adding the new event. This preserves the existing clear-on-commit-failure behavior. Does that overflow behavior match the intended policy? I have not pushed the local change because Maven stops at the enforcer: this checkout requires JDK [17,18), but the host has JDK 26.0.2.1. Let's keep this PR to the leak: clear in the outer finally, plus a one-time status-logger warning when a batch is discarded, like the writeWhileNotRunningLogged flag in #4246. Capping the buffer means deciding which event dies and probably a new attribute, so open a separate issue for it. Also drop the duplicate-rows wording: the connection runs with autoCommit false, so a failed commit rolls back and nothing lands. -- 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]
