[
https://issues.apache.org/jira/browse/HBASE-13919?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14702469#comment-14702469
]
Lars Hofhansl commented on HBASE-13919:
---------------------------------------
When you say "Connecting to cluster" takes 20 mins, that's when everything is
down including ZK, right? I observed that too, it's somewhat of a worst case,
where we timeout on ZK and the HMaster... repeatedly.
There are also server-server timeouts and client-server timeouts. I'm still
planning to document all these and make sense of them.
> Rationalize Client Timeout
> --------------------------
>
> Key: HBASE-13919
> URL: https://issues.apache.org/jira/browse/HBASE-13919
> Project: HBase
> Issue Type: Brainstorming
> Reporter: Lars Hofhansl
>
> There are current many setting that influence how/when an HBase client times
> out.
> This is hard to configure, hard to understand, and badly documented.
> Instead there should as few tweaks as needed, and their setting should make
> sense.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
