[
https://issues.apache.org/jira/browse/HIVE-8901?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14217556#comment-14217556
]
Lefty Leverenz commented on HIVE-8901:
--------------------------------------
Doc note: This changes the defaults for configuration parameters
*hive.hmshandler.retry.attempts* and *hive.hmshandler.retry.interval* in
HiveConf.java. They were introduced in Hive 0.10.0 by HIVE-3484 but aren't
documented in the wiki yet. So the Metastore section of Configuration
Properties needs to document them with the original defaults and new values.
* [Configuration Properties -- MetaStore |
https://cwiki.apache.org/confluence/display/Hive/Configuration+Properties#ConfigurationProperties-MetaStore]
By the way, *hive.hmshandler.force.reload.conf* also needs to be documented
(HIVE-3484).
> increase retry attempt, interval on metastore database errors
> -------------------------------------------------------------
>
> Key: HIVE-8901
> URL: https://issues.apache.org/jira/browse/HIVE-8901
> Project: Hive
> Issue Type: Bug
> Reporter: Thejas M Nair
> Assignee: Thejas M Nair
> Labels: TODOC15
> Fix For: 0.15.0
>
> Attachments: HIVE-8901.1.patch
>
>
> The current defaults of hive.hmshandler.retry.attempts=1 and
> hive.hmshandler.retry.interval=1s are very small. In some cases one retry is
> not sufficient for successful command execution.
>
> I have seen cases, specially after metastore being idle for sometime, when it
> takes 2 attempts for the db action to succeed. In this case, it as a
> Communications link failure/connection lost error.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
