[
https://issues.apache.org/jira/browse/HBASE-13348?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14386828#comment-14386828
]
stack commented on HBASE-13348:
-------------------------------
Looks good to me. +1. On commit put a '.' between meta and serverops so
'meta.serverop' but this is nit.
> Separate the thread number configs for meta server and server operations
> ------------------------------------------------------------------------
>
> Key: HBASE-13348
> URL: https://issues.apache.org/jira/browse/HBASE-13348
> Project: HBase
> Issue Type: Improvement
> Components: master
> Reporter: Liu Shaohui
> Assignee: Liu Shaohui
> Priority: Minor
> Fix For: 2.0.0
>
> Attachments: HBASE-13348-v1.diff
>
>
> Currently, the config keys for thread number of meta server and server
> operations in HMaster are same:
> See: HMaster.java #993
> {code}
> this.service.startExecutorService(ExecutorType.MASTER_SERVER_OPERATIONS,
> conf.getInt("hbase.master.executor.serverops.threads", 5));
>
> this.service.startExecutorService(ExecutorType.MASTER_META_SERVER_OPERATIONS,
> conf.getInt("hbase.master.executor.serverops.threads", 5));
> {code}
> In large cluster, we usually enlarge the thread number for server operation
> separately to make the master handle regionserver shutdown events quickly
> in some extremely cases.
> So I think we need separate the thread number config for the two operations.
> Suggestions are welcomed. Thanks~
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
