wy96f commented on a change in pull request #2845: ARTEMIS-2336 Use zero copy
to replicate journal/page/large message file (AGAIN)
URL: https://github.com/apache/activemq-artemis/pull/2845#discussion_r327948683
##########
File path:
artemis-core-client/src/main/java/org/apache/activemq/artemis/core/remoting/impl/netty/NettyConnection.java
##########
@@ -355,12 +373,24 @@ private boolean canWrite(final int requiredCapacity) {
return canWrite;
}
- private Object getFileObject(RandomAccessFile raf, FileChannel fileChannel,
long offset, int dataSize) {
- if (channel.pipeline().get(SslHandler.class) == null) {
- return new NonClosingDefaultFileRegion(fileChannel, offset, dataSize);
+ private Object getFileObject(FileChannel fileChannel, long offset, int
dataSize) {
+ if (USE_FILE_REGION && channel.pipeline().get(SslHandler.class) == null)
{
Review comment:
> So it seems that Netty is slowing sending them...can see 2 reasons for
that:
>
> * we are sending too much for real
> * the receiving broker is not reading them fast enough
>
> I cannot find other reasons, other insights?
Actually broker was not sending built up data in the queue after some point
until timeout failure. It seems watermark mechanism triggered some bug
resulting in not resuming transfer in `ChunkedWriteHandler`.
----------------------------------------------------------------
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:
[email protected]
With regards,
Apache Git Services
