[
https://issues.apache.org/jira/browse/QPID-8305?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16830066#comment-16830066
]
ASF GitHub Bot commented on QPID-8305:
--------------------------------------
overmeulen commented on pull request #30: QPID-8305: [Broker-J][JDBC Message
Store] Performance regression when increasing the number of queues linked to a
topic
URL: https://github.com/apache/qpid-broker-j/pull/30
With this fix I observed a 20% improvement of the overall throughput when I
have 1 topic with 5 queues (# binding key for all of them).
Note that there is most likely another bottleneck somewhere else: after this
fix I see a 50% regression of the overall throughput between 1 and 5 queues
whereas at the database level I only have a 20% increase of the elapsed time...
----------------------------------------------------------------
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]
> [Broker-J][JDBC Message Store] Performance regression when increasing the
> number of queues linked to a topic
> ------------------------------------------------------------------------------------------------------------
>
> Key: QPID-8305
> URL: https://issues.apache.org/jira/browse/QPID-8305
> Project: Qpid
> Issue Type: Improvement
> Components: Broker-J
> Affects Versions: qpid-java-broker-7.1.0
> Reporter: Olivier VERMEULEN
> Priority: Major
>
> When sending a message to a topic, if this message is routed to N queues then
> we will do N+2 database inserts: 1 for the metadata, 1 for the content and 1
> for each message/queue mapping (QPID_QUEUE_ENTRIES). The last N inserts could
> be batched in a single operation.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
