[
https://issues.apache.org/jira/browse/HBASE-3110?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
stack updated HBASE-3110:
-------------------------
Attachment: 3110.txt
So, this test is failing because it decides of a sudden to connect to a
different zk cluster, not the one started by this test. When it does this it
gets bad master -- a stale master address -- and so fails.
It of a sudden connects to a stale ensemble because it resorts to using default
location for zk ensemble. Its using a connection that hasn't been updated with
the actual location of the zk ensemble. Because we're leaking a zk cluster in
the tests somewhere -- some test is starting it up but not shutting it down --
all subsequent tests fail to start an ensemble on the default and resort to
unorthodox ports.
I'm looking for the culprit but haven't figured it yet.
> TestReplicationSink failing in TRUNK up on Hudson
> -------------------------------------------------
>
> Key: HBASE-3110
> URL: https://issues.apache.org/jira/browse/HBASE-3110
> Project: HBase
> Issue Type: Bug
> Reporter: stack
> Fix For: 0.90.0
>
> Attachments: 3110.txt
>
>
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
