[ https://issues.apache.org/jira/browse/ARTEMIS-350?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15158932#comment-15158932 ]
ASF GitHub Bot commented on ARTEMIS-350: ---------------------------------------- GitHub user jbertram opened a pull request: https://github.com/apache/activemq-artemis/pull/400 ARTEMIS-350 fix for potential race It's possible for the latch used for flow control here to get out of sync. In other words, multiple count-downs can occur between count-ups so that the latch always has a count > 0. When this situation arises then every single packet sent to the replica is delayed by 5 seconds. The solution here essentially is to reset the latch back to zero every time instead of decrementing it by one with a normal count-down. This should still maintain overall flow-control, but also avoid this race condition. You can merge this pull request into a Git repository by running: $ git pull https://github.com/jbertram/activemq-artemis ARTEMIS-350 Alternatively you can review and apply these changes as the patch at: https://github.com/apache/activemq-artemis/pull/400.patch To close this pull request, make a commit to your master/trunk branch with (at least) the following in the commit message: This closes #400 ---- commit 366316352f1e0169effb1b455832a9188aed671c Author: jbertram <jbert...@apache.org> Date: 2016-02-16T16:45:25Z ARTEMIS-350 fix for potential race It's possible for the latch used for flow control here to get out of sync. In other words, multiple count-downs can occur between count-ups so that the latch always has a count > 0. When this situation arises then every single packet sent to the replica is delayed by 5 seconds. The solution here essentially is to reset the latch back to zero every time instead of decrementing it by one with a normal count-down. This should still maintain overall flow-control, but also avoid this race condition. ---- > possible OOM in replication manager > ----------------------------------- > > Key: ARTEMIS-350 > URL: https://issues.apache.org/jira/browse/ARTEMIS-350 > Project: ActiveMQ Artemis > Issue Type: Bug > Reporter: Andy Taylor > Assignee: Andy Taylor > Fix For: 1.3.0 > > > This is becausee we have no flow control, we need to throttle back when the > channel is not writable -- This message was sent by Atlassian JIRA (v6.3.4#6332)