jbertram commented on issue #2850: ARTEMIS-2504 implement retroactive addresses
URL: https://github.com/apache/activemq-artemis/pull/2850#issuecomment-539140540
   > if an address is paging to disk, ring queue still gets new messages 
delivered, theyre just paged references.
   According to my tests the ring queue will *not* actually get those messages. 
From what I can tell messages are paged at the address level, and they aren't 
added to the actual queue until they are de-paged.
   Regarding the performance issues, my initial recommendation would be to 
limit the size of the ring queue to avoid such problems. As for your ideas...
   > is there a way to control the copy over so its not done as one big bang 
(e.g. as fast as broker can possible do), but feed the new queue at the rate 
the consumer is consuming there for avoiding the very big spike in object 
   I understand your point about throttling the copy speed, but I'm not sure 
that will provide the semantics a consumer expects. Once the queue is created 
then it will receive any "new" message sent to the queue's address (at least in 
the multicast use-case). If the "old," retroactive messages are not all 
immediately added to the queue then the consumer might receive "new" messages 
before it receives all the "old" ones. 
   Also, since messages are added to the queue when it is first created and 
before the client even consumes anything how can we calculate the consumption 
rate in order to throttle the copy rate?
   > Is it possible to make it so that we dont have to copy, but it simply is a 
message ref?
   I'll explore this possibility.

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:

With regards,
Apache Git Services

Reply via email to