[
https://issues.apache.org/jira/browse/HBASE-24436?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17117263#comment-17117263
]
Andrew Kyle Purtell edited comment on HBASE-24436 at 5/27/20, 3:09 AM:
-----------------------------------------------------------------------
Repeating my question from the PR: How is it more efficient if scheduled at the
region level? The goal as o understand it is to stop a large store with lots of
files from hogging resources but if you don’t use a special executor or
priority queue to round robin the open work over the families then the
submitted runnables will burst in the same way when the work for a particular
family is scheduled.
Also, the resource is “store file opener” so IMHO belongs in Store (or HStore),
not Region, although that is a much more minor issue.
was (Author: apurtell):
Repeating my question from the PR: How is it more efficient if scheduled at the
region level? The goal as o understand it is to stop a large store with lots of
files from hogging resources but if you don’t use a special executor or
priority queue to round robin the open work then the submitted runnables will
burst in the same way.
Also, the resource is “store file opener” so IMHO belongs in Store (or HStore),
not Region, although that is a much more minor issue.
> The store file open and close thread pool should be shared at the region level
> ------------------------------------------------------------------------------
>
> Key: HBASE-24436
> URL: https://issues.apache.org/jira/browse/HBASE-24436
> Project: HBase
> Issue Type: Improvement
> Reporter: Junhong Xu
> Assignee: Junhong Xu
> Priority: Minor
>
> For now, we provide threads per column family evenly in general, but there
> are some cases that some column families have much more store files than
> others( maybe that's the life, right? ). So in that case, some Stores have
> beed done quickly while others are struggling.We should share the thread pool
> at the region level in case of data skew.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
