michaelandrepearce commented on a change in pull request #2707: ARTEMIS-2002
Proton transport objects leaked
URL: https://github.com/apache/activemq-artemis/pull/2707#discussion_r295663234
##########
File path:
artemis-protocols/artemis-amqp-protocol/src/main/java/org/apache/activemq/artemis/protocol/amqp/proton/AMQPConnectionContext.java
##########
@@ -389,13 +408,15 @@ public void onRemoteOpen(Connection connection) throws
Exception {
initialise();
/*
- * This can be null which is in effect an empty map, also we really don't
need to check this for in bound connections
- * but its here in case we add support for outbound connections.
- * */
+ * This can be null which is in effect an empty map, also we really
don't need to check this for in bound connections
+ * but its here in case we add support for outbound connections.
+ * */
if (connection.getRemoteProperties() == null ||
!connection.getRemoteProperties().containsKey(CONNECTION_OPEN_FAILED)) {
long nextKeepAliveTime = handler.tick(true);
- if (nextKeepAliveTime != 0 && scheduledPool != null) {
- scheduledPool.schedule(new ScheduleRunnable(), (nextKeepAliveTime
- TimeUnit.NANOSECONDS.toMillis(System.nanoTime())), TimeUnit.MILLISECONDS);
+ synchronized (schedulingLock) {
Review comment:
I have a very big concern here. Locks in connection stuff, artemis was built
on the premise of being nonblocking. As time goes on more and more blocking
locks are being added to core code paths. That then take months unravellling
later. Can we please spend the time upfront now to find a solution to avoid
adding sync locks.
----------------------------------------------------------------
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
