apurtell commented on a change in pull request #1945:
URL: https://github.com/apache/hbase/pull/1945#discussion_r445735442
##########
File path:
hbase-zookeeper/src/main/java/org/apache/hadoop/hbase/zookeeper/ZKWatcher.java
##########
@@ -79,8 +86,22 @@
// listeners to be notified
private final List<ZKListener> listeners = new CopyOnWriteArrayList<>();
+ // Single threaded executor pool that processes event notifications from
Zookeeper. Events are
+ // processed in the order in which they arrive (pool backed by an unbounded
fifo queue). We do
+ // this to decouple the event processing from Zookeeper's ClientCnxn's
EventThread context.
+ // EventThread internally runs a single while loop to serially process all
the events. When events
+ // are processed by the listeners in the same thread, that blocks the
EventThread from processing
+ // subsequent events. Processing events in a separate thread frees up the
event thread to continue
Review comment:
I see, this was at the root of the test failures.
----------------------------------------------------------------
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]
