[
https://issues.apache.org/jira/browse/HBASE-27091?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Yu Li resolved HBASE-27091.
---------------------------
Hadoop Flags: Reviewed
Release Note: Loading table descriptors with multiple threads is supported
after HBASE-27091 (but turned off by default). The thread number could be
configured via `hbase.tabledescriptor.parallel.load.threads`, and setting it
back to zero will turn off the feature.
Resolution: Implemented
Merged into master via c6298c7.
Thanks [~heliangjun] for the contribution and [~huaxiangsun] for the review.
> Speed up the loading of table descriptor from filesystem
> --------------------------------------------------------
>
> Key: HBASE-27091
> URL: https://issues.apache.org/jira/browse/HBASE-27091
> Project: HBase
> Issue Type: Improvement
> Components: master
> Affects Versions: 3.0.0-alpha-3
> Reporter: LiangJun He
> Assignee: LiangJun He
> Priority: Minor
> Fix For: 3.0.0-alpha-4
>
>
> If there are a large number of tables in the HBase cluster, it will take a
> long time to fully load the table descriptor from filesystem for the first
> time.
> In our production cluster, there were 50000 + tables. It took several minutes
> to load the table descriptor from filesystem. This problem seriously affects
> the performance of HMaster active/standby switchover.
> We should support concurrent loading to solve this problem.
--
This message was sent by Atlassian Jira
(v8.20.7#820007)
