[
https://issues.apache.org/jira/browse/HBASE-8483?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13649885#comment-13649885
]
stack commented on HBASE-8483:
------------------------------
[~ericyu] Thanks for having a go at this.
Usually we synchronize on
synchronized (masterAndZKLock) {
... for example in the abort method when doing zk operations but then on the
way out, you take a synchronization on this:
+ synchronized(this) {
+ closeZooKeeperWatcher();
What is your thinking?
Else patch looks good. Nice test.
> HConnectionManager can leak ZooKeeper connections when using
> deleteStaleConnection
> ----------------------------------------------------------------------------------
>
> Key: HBASE-8483
> URL: https://issues.apache.org/jira/browse/HBASE-8483
> Project: HBase
> Issue Type: Bug
> Components: Client
> Affects Versions: 0.94.4
> Reporter: Eric Yu
> Priority: Critical
> Fix For: 0.95.1
>
> Attachments: HBASE-8483-0.94.patch, HBASE-8483.patch,
> LeakZKConnections.java
>
>
> If one thread calls deleteStaleConnection while other threads are using
> connection, can leak ZK connections.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
