You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@hudi.apache.org by "sivabalan narayanan (Jira)" <ji...@apache.org> on 2021/10/21 17:18:00 UTC
[jira] [Updated] (HUDI-2595) Guard all writes to metadata table
even for tables where locking is not enabled
[ https://issues.apache.org/jira/browse/HUDI-2595?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
sivabalan narayanan updated HUDI-2595:
--------------------------------------
Fix Version/s: 0.10.0
> Guard all writes to metadata table even for tables where locking is not enabled
> -------------------------------------------------------------------------------
>
> Key: HUDI-2595
> URL: https://issues.apache.org/jira/browse/HUDI-2595
> Project: Apache Hudi
> Issue Type: Improvement
> Reporter: sivabalan narayanan
> Assignee: sivabalan narayanan
> Priority: Blocker
> Fix For: 0.10.0
>
>
> Guard all writes to metadata table even for tables where locking is not enabled.
> As of now, we guard all writes to metadata table by taking a data table lock. This might work well for tables where multi-writer is enabled and so locking does take place. but for tables where multi-writer is not enabled, changes to metadata table happens w/o any lock.
> so, if async compaction is running, it could potentially overstep w/ just 1 writer. both could concurrently try to update metadata table.
>
>
--
This message was sent by Atlassian Jira
(v8.3.4#803005)