[
https://issues.apache.org/jira/browse/HIVE-23349?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Denys Kuzmenko updated HIVE-23349:
----------------------------------
Description:
2 concurrent MERGE INSERT operations generate duplicates due to lack of
locking.
MERGE INSERT is treated as regular INSERT, it acquires SHARED_READ lock that
doesn't prevent other SHARED_READs. We should use EXCLUSIVE lock here or
EXCL_WRITE if hive.txn.write.xlock=false;
{code}
create table target (a int, b int) stored as orc TBLPROPERTIES
('transactional'='true')");
insert into target values (1,2), (3,4)
create table source (a int, b int)
{code}
execute in parallel:
{code}
insert into source values (5,6), (7,8)
{code}
was:
2 concurrent MERGE INSERT operations generate duplicates due to lack of
locking.
MERGE INSERT is treated as regular INSERT, it acquires SHARED_READ lock that
doesn't prevent other SHARED_READs. We should use EXCLUSIVE lock here or
EXCL_WRITE if hive.txn.write.xlock=false (INSERT would acquire SHARED_WRITE in
this case);
{code}
create table target (a int, b int) stored as orc TBLPROPERTIES
('transactional'='true')");
insert into target values (1,2), (3,4)
create table source (a int, b int)
{code}
execute in parallel:
{code}
insert into source values (5,6), (7,8)
{code}
> ACID: Concurrent MERGE INSERT operations produce duplicates
> -----------------------------------------------------------
>
> Key: HIVE-23349
> URL: https://issues.apache.org/jira/browse/HIVE-23349
> Project: Hive
> Issue Type: Bug
> Components: Transactions
> Reporter: Denys Kuzmenko
> Assignee: Denys Kuzmenko
> Priority: Major
>
> 2 concurrent MERGE INSERT operations generate duplicates due to lack of
> locking.
> MERGE INSERT is treated as regular INSERT, it acquires SHARED_READ lock that
> doesn't prevent other SHARED_READs. We should use EXCLUSIVE lock here or
> EXCL_WRITE if hive.txn.write.xlock=false;
> {code}
> create table target (a int, b int) stored as orc TBLPROPERTIES
> ('transactional'='true')");
> insert into target values (1,2), (3,4)
> create table source (a int, b int)
> {code}
> execute in parallel:
> {code}
> insert into source values (5,6), (7,8)
> {code}
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
