franz1981 commented on issue #2479: ARTEMIS-2211 Refactor ByteBuffer pooling, 
alignment and zeroing
URL: https://github.com/apache/activemq-artemis/pull/2479#issuecomment-470456911
 
 
   @michaelandrepearce 
   > also whats occuring with the possible leakage has that been addressed
   
   We have 2 leaks with NIO/MAPPED seq factories:
   
   1. on compaction: it has been addressed not in this PR by NOT using 
`ByteBuffer` pooling ie it will allocate upfront and free the buffers when 
finished
   2. on paging/while writing on the journal
   
   The second one is due to thread-local `ByteBuffer` pooling + our thread pool 
+ type of load: 
   
   - we have burst of journal activity that will make N threads to be created 
   - the burst of load will finish and more then 1 minute has passed: many 
threads got disposed, but the thread-pooled `ByteBuffer` will be cleaned up 
only if a GC will happen, but suppose it won't happen
   - another burst of journal activity will make other N threads to be created
   - we ends up having many `ByteBuffer` instances alive: old ones + new ones

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