You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues-all@impala.apache.org by "ASF subversion and git services (Jira)" <ji...@apache.org> on 2021/08/06 01:35:00 UTC
[jira] [Commented] (IMPALA-10817) Share metastoreHmsDDL lock b/w
CatalogOpExecutor and Catalog metastore server
[ https://issues.apache.org/jira/browse/IMPALA-10817?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17394415#comment-17394415 ]
ASF subversion and git services commented on IMPALA-10817:
----------------------------------------------------------
Commit 37758f0c0c50e0bf343b923fe7bc357375ed89c2 in impala's branch refs/heads/master from Sourabh Goyal
[ https://gitbox.apache.org/repos/asf?p=impala.git;h=37758f0 ]
IMPALA-10817: Share metastoreHmsDDL lock b/w CatalogOpExecutor and
Catalog metastore server
Currently, when doing create/drop table/db from catalogD,
catalogOpExecutor (via Impala Shell) and metastore server (via HS2)
acquires lock on their own lock objects to prevent concurrent
create/drop operations in HMS. But that does not prevent these
concurrent operations across CatalogOpExecutor and Metastore server.
For example currently a user can perform create/drop HMS operation
from Impala shell and catalog metastore server concurrently which is
not the desired behavior.
This patch unifies the lock object which will be acquired by both
catalogOpExecutor as well as metastore server.
Testing:
Relying on existing tests since it is a small refactoring.
Change-Id: I60d4f3a49eb843fa8640cd21d623fd8dda770001
Reviewed-on: http://gerrit.cloudera.org:8080/17713
Tested-by: Impala Public Jenkins <im...@cloudera.com>
Reviewed-by: <ki...@cloudera.com>
Reviewed-by: Vihang Karajgaonkar <vi...@cloudera.com>
> Share metastoreHmsDDL lock b/w CatalogOpExecutor and Catalog metastore server
> -----------------------------------------------------------------------------
>
> Key: IMPALA-10817
> URL: https://issues.apache.org/jira/browse/IMPALA-10817
> Project: IMPALA
> Issue Type: Improvement
> Components: Catalog
> Reporter: Sourabh Goyal
> Priority: Major
>
> Currently, when doing create/drop table/db from catalogD, catalogOpExecutor (via Impala Shell) and metastore server (via HS2) acquires lock on their own respective lock objects to prevent concurrent create/drop operations in HMS. But that does not prevent these concurrent operations across CatalogOpExecutor and Metastore server. For example currently a user can perform create/drop HMS operation from Impala shell and catalog metastore server concurrently which is not the desired behavior
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-all-unsubscribe@impala.apache.org
For additional commands, e-mail: issues-all-help@impala.apache.org