[
https://issues.apache.org/jira/browse/HBASE-13351?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14484547#comment-14484547
]
Mikhail Antonov commented on HBASE-13351:
-----------------------------------------
The practical (in our case) small difference between the two is that
callExecutor uses bounded priority array-based queue, and priority and
replicator executors are using linked blocking queues, and also with default
configs callExecutor would have 1 backing queue (when 2 handlers are in use),
and priority executor is hardcoded to use 2. Which probably doesn't really
matter.
Seems like in this case we're just happen to hit corner case in a sense that
normally in prod cluster one expects user rpcs to be the prevailing type of
calls, and in this test it's just not the case (btw..the same effect would
apply when making admin-invoked calls to use high-priority thread pool, since
inside minicluster..well, everyone is admin :)).
> Annotate internal MasterRpcServices methods with admin priority
> ---------------------------------------------------------------
>
> Key: HBASE-13351
> URL: https://issues.apache.org/jira/browse/HBASE-13351
> Project: HBase
> Issue Type: Improvement
> Components: master
> Reporter: Josh Elser
> Assignee: Josh Elser
> Fix For: 2.0.0, 1.1.0
>
> Attachments: HBASE-13351-v1.patch, HBASE-13351.patch
>
>
> HBASE-12071, among other things, introduced annotating RPC methods to give
> certain methods priority over others. Namely, this helps ensure that client
> requests cannot starve out internal RPC between master and regionserver.
> Similarly, we can do the same thing for Master RPC methods that are invoked
> by RS's.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
