[
https://issues.apache.org/jira/browse/HBASE-3808?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13037422#comment-13037422
]
stack commented on HBASE-3808:
------------------------------
@Brock Yes. There is a little bit of a convention naming threads that you
might want to pick up on. Here are some illustartions of the pattern:
Here is how we name the master where MASTER=='master'
{code}
// set the thread name now we have an address
setName(MASTER + "-" + this.serverName.toString());
{code}
Note how we get the 'servername' in there. Helps in for example a testing
context where there might be N masters all running in the same JVM distingushed
by port number only; having the hosting server name as prefix on the thread can
help debugging.
Here is how the catalogjanitor that runs in each master is named:
{code}
server.getServerName() + "-CatalogJanitor"
{code}
The bulkassigner that runs in each master:
{code}
return this.server.getServerName() + "-" + this.getClass().getName();
{code}
That kinda thing.
Not sure how you will distingush between the instances of Executors? Perhaps
there is a increment of some time you can add to the name?
While you are at it, you might want to add toString to the handlers in
regionserver.
Thank you for taking a look.
Beware: this might be a silly enterprise, perhaps not doable. It was just a
notion I had.
> Implement Executor.toString for master handlers at least
> --------------------------------------------------------
>
> Key: HBASE-3808
> URL: https://issues.apache.org/jira/browse/HBASE-3808
> Project: HBase
> Issue Type: Bug
> Reporter: stack
> Priority: Minor
>
> On shutdown, if still outstanding Executors queued then when ExecutorService
> lists what is outstanding, the list will be other than a list of default
> toString implementations of ServerShutdownHandler objects.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
