[
https://issues.apache.org/jira/browse/HBASE-21333?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16659630#comment-16659630
]
stack commented on HBASE-21333:
-------------------------------
HBASE-21237 removed the batching of open/close RPCs in branch-2.0 and
branch-2.1 because it doesn't do errors properly. Testing with HBASE-21237
temporarily backed-out, we do about 90 assigns a second.
> [amv2] large cluster startup is slow
> ------------------------------------
>
> Key: HBASE-21333
> URL: https://issues.apache.org/jira/browse/HBASE-21333
> Project: HBase
> Issue Type: Sub-task
> Components: amv2
> Affects Versions: 2.1.0
> Reporter: stack
> Priority: Major
> Attachments: 2.1.1.129578.alloc.svg, 2.1.1.129578.cpu.svg,
> 2.1.1.129578.lock.svg
>
>
> Testing startup of cluster with 500+ nodes and .5M regions takes a few hours.
> This is a 2.1.x cluster with batching disabled.
> Looking at what the Master is doing, its mostly just parsing regionserver
> reports.
> Stats to follow.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
