[
https://issues.apache.org/jira/browse/HIVE-29789?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18101366#comment-18101366
]
Dayakar M commented on HIVE-29789:
----------------------------------
[~zabetak] No [HIVE-29762|https://issues.apache.org/jira/browse/HIVE-29762] fix
will not be increasing the chances for dataloss. The fix eliminates/minimizes
the hash collision issue. The initial JIRA
[HIVE-24428|https://issues.apache.org/jira/browse/HIVE-24428] addresses the
issue for a single HMS instance, for multiple HMS instance case still there can
be a chance of data loss. That needs to be checked and provide a proper fix so
raised the current JIRA.
> Concurrent add_partitions requests to multiple HMS instances may lead to data
> loss
> ----------------------------------------------------------------------------------
>
> Key: HIVE-29789
> URL: https://issues.apache.org/jira/browse/HIVE-29789
> Project: Hive
> Issue Type: Bug
> Components: Metastore
> Reporter: Dayakar M
> Assignee: Dayakar M
> Priority: Major
>
> Concurrent add_partitions requests to multiple HMS instances may lead to data
> loss.
> It's a followup Jira for
> [HIVE-24428|https://issues.apache.org/jira/browse/HIVE-24428] which fixed the
> issue by providing synchronization support for a single HMS instance but same
> issue can be there if multiple HMS instances are involved. So this has to be
> cheked properly and need a proper fix.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
