[
https://issues.apache.org/jira/browse/HIVE-12164?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14962208#comment-14962208
]
Lefty Leverenz commented on HIVE-12164:
---------------------------------------
Doc note: This revises the configuration parameter *hive.stats.dbclass* and
removes five others (*hive.stats.jdbcdriver*, *hive.stats.dbconnectionstring*,
*hive.stats.jdbc.timeout*, *hive.stats.retries.max*, and
*hive.stats.retries.wait*) so they need to be updated with version information
in the Statistics section of the Configuration Properties wikidoc.
Also, one example in the Statistics wikidoc needs version information.
* [Statistics in Hive -- Usage -- Newly Created Tables |
https://cwiki.apache.org/confluence/display/Hive/StatsDev#StatsDev-NewlyCreatedTables]
* [Configuration Properties -- Statistics |
https://cwiki.apache.org/confluence/display/Hive/Configuration+Properties#ConfigurationProperties-Statistics]
> Remove jdbc stats collection mechanism
> --------------------------------------
>
> Key: HIVE-12164
> URL: https://issues.apache.org/jira/browse/HIVE-12164
> Project: Hive
> Issue Type: Task
> Components: Statistics
> Affects Versions: 2.0.0
> Reporter: Ashutosh Chauhan
> Assignee: Ashutosh Chauhan
> Labels: TODOC2.0
> Fix For: 2.0.0
>
> Attachments: HIVE-12164.1.patch, HIVE-12164.2.patch,
> HIVE-12164.3.patch, HIVE-12164.patch
>
>
> Though there are some deployments using it, usually its painful to setup
> since a valid hive-site.xml is needed on all task nodes (containing
> connection details) and for large tasks (with thousands of tasks) results in
> a scalability issue with all of them hammering DB at nearly same time.
> Because of these pain points alternative stats collection mechanism were
> added. FS stats based system is default for some time.
> We should remove jdbc stats collection mechanism as it needlessly adds
> complexity in TS and FS operators w.r.t key handling.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
