franz1981 commented on issue #2845: ARTEMIS-2336 Use zero copy to replicate journal/page/large message file (AGAIN) URL: https://github.com/apache/activemq-artemis/pull/2845#issuecomment-541495785 @wy96f sadly the rate limiter algorithm in guava is not very good fore fast paced networks and if with small token size, despite what the doc says about it :) Netty is considering 0 sized the FileRegion sent on the outbound buffer afaik and just "unknown" the one based on chunked files, but the point is that the latter will be accumulated and won't back-pressure the channel.becauee are sent through a background accumulator (the additional outbound channel in the pipeline) ie the solution to provide a custom estimator makes sense imho, but I need my computer at hand to try it. My point on the existing sync done is that is not actually correct because it assume that the previous files are being sent before sending that sync done while is not in any case, because Netty is asynchronous...the chunked file case just make it more evident.. My 2 cents :)
---------------------------------------------------------------- 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. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services