[
https://issues.apache.org/jira/browse/HBASE-13919?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14588651#comment-14588651
]
Lars Hofhansl commented on HBASE-13919:
---------------------------------------
Here are some of the settings:
* hbase.client.retries.number
* hbase.client.pause
* hbase.rpc.timeout
* hbase.client.scanner.timeout.period
* hbase.client.operation.timeout
* hbase.client.meta.operation.timeout
* hbase.rpc.shortoperation.timeout
* zookeeper.recovery.retry
* zookeeper.session.timeout
* zookeeper.recovery.retry.intervalmill
It's hard to understand what all of these mean and how they interact. I'm going
to start by documenting all of these (unless they are already), and then see
whether we actually need so many option, or whether we can derive some settings
from the others.
Also it seems the defaults still lead to 20m overall timeout when ZK and HBase
servers are down.
> Rationaliize 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)
