bbeaudreault commented on PR #5350: URL: https://github.com/apache/hbase/pull/5350#issuecomment-1679174377
@Apache9 I just pushed a change which includes a bit of a refactor: - Change NettyUnsafeUtils.closeDirect to `closeImmediately`, and have it manage SO_LINGER there. Improve javadoc to say why this is important. - Move all of the writability protections (aborting and setAutoRead) into a new single ChannelDuplexHandler. I felt it was awkward to have this in the Connection, where we need to assert inEventLoop.... Feels better to just put the code itself in the event loop. - Added tests. It was hard to trigger the fatal threshold. In order to do that I had to change NettyServerCall `sendResponseIfReady` to call NettyServerRpcConnection `doRespond`, so that I could mock the writeAndFlush to skip the flush. The side effect here is now that we do safeWriteAndFlush, we will log any failures from normal responses just like other callers to doRespond. Let me know what you think -- 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]
