[
https://issues.apache.org/jira/browse/ACCUMULO-1586?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Michael Berman updated ACCUMULO-1586:
-------------------------------------
Fix Version/s: 1.6.0
1.5.1
Status: Patch Available (was: Open)
Chose the "create the timed processor outside the port binding loop" option.
Kept a TServerUtils.startTServer() delegate that still takes the unwrapped
Processor so other consumers don't have to wrap it themselves if they're not
doing anything fancy.
> If initial port binding fails, ThriftMetrics MBean is never registered for
> subsequently bound TServer
> -----------------------------------------------------------------------------------------------------
>
> Key: ACCUMULO-1586
> URL: https://issues.apache.org/jira/browse/ACCUMULO-1586
> Project: Accumulo
> Issue Type: Bug
> Affects Versions: 1.5.0
> Reporter: Michael Berman
> Assignee: Michael Berman
> Fix For: 1.5.1, 1.6.0
>
> Attachments: ACCUMULO-1586.patch
>
>
> When trying a range of ports, the MBean for reporting metrics is registered
> before the connection has had an opportunity to fail. If it fails and we try
> a new port, the same MBean objectName is used for registering the new
> ThriftMetrics object created for the new server. This fails because there is
> already an MBean registered with that name. A couple possible fixes: wait to
> register until we're sure the connection has succeeded; create the timed
> processor outside the port-binding loop; or deregister the bean after a
> failure.
--
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
