[
https://issues.apache.org/jira/browse/IMPALA-12787?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17815210#comment-17815210
]
ASF subversion and git services commented on IMPALA-12787:
----------------------------------------------------------
Commit 68e61c1aabacee5da37b1f66842d8a2354d0e215 in impala's branch
refs/heads/master from Zoltan Borok-Nagy
[ https://gitbox.apache.org/repos/asf?p=impala.git;h=68e61c1aa ]
IMPALA-12787: Concurrent DELETE and UPDATE operations on Iceberg tables can be
problematic
If an UPDATE operation runs concurrently with a DELETE operation, and
the DELETE commits first, then the UPDATE can revive deleted rows. This
is because only RowDelta.validateNoConflictingDataFiles() is called, but
RowDelta.validateNoConflictingDeleteFiles() is not. Therefore, the
UPDATE operation ignores the concurrently written delete files.
This patch adds RowDelta.validateNoConflictingDeleteFiles() to UPDATE
operations.
Testing
* added a stress test to validate concurrent DELETE and UPDATE
operations
Change-Id: I9e581ea17fa8f6ccd9c87aaad1281bb694079f6e
Reviewed-on: http://gerrit.cloudera.org:8080/20999
Reviewed-by: Impala Public Jenkins <[email protected]>
Tested-by: Impala Public Jenkins <[email protected]>
> Concurrent DELETE and UPDATE operations on Iceberg tables can be problematic
> ----------------------------------------------------------------------------
>
> Key: IMPALA-12787
> URL: https://issues.apache.org/jira/browse/IMPALA-12787
> Project: IMPALA
> Issue Type: Bug
> Components: Catalog
> Reporter: Zoltán Borók-Nagy
> Assignee: Zoltán Borók-Nagy
> Priority: Major
> Labels: impala-iceberg
>
> If an UPDATE operation runs concurrently with a DELETE operation, and the
> DELETE commits first, then the UPDATE can revive deleted rows.
> This is because only RowDelta.validateNoConflictingDataFiles() is called, but
> RowDelta.validateNoConflictingDeleteFiles() is not. Therefore, the UPDATE
> operation ignores the concurrently written delete files.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
