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-535803060
 
 
   If both cases (with/without file region) are saturating the network, why the 
latter will take more time? The total amount of data sent should be the same...
   Do you have tried master as well? 
   I'm start to think that the pipelining happening while copying data in a 
non-netty thread and a separate Netty thread sending them across network is 
beneficial to improve the overall throughput, because we can do something (ie 
reading the file) while Netty is taking care to send data across network. But 
that means that if we don't use file regions I expect that network is not 
saturated 100% of the time and sometime wait ChunkInput to finish reading data 
and saturate it again...
   I don't know if your acceptor configuration of TCP buffer is working,need to 
inspect the logs...
   And profiling could be helpful as well...

----------------------------------------------------------------
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

Reply via email to