[
https://issues.apache.org/jira/browse/HIVE-21766?focusedWorklogId=249088&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-249088
]
ASF GitHub Bot logged work on HIVE-21766:
-----------------------------------------
Author: ASF GitHub Bot
Created on: 28/May/19 03:09
Start Date: 28/May/19 03:09
Worklog Time Spent: 10m
Work Description: maheshk114 commented on pull request #647: HIVE-21766 :
Select * returns no rows in hive bootstrap from a statior dynamic partitioned
managed table with Timestamp type as partition column from on prem to WASB even
though count ( * ) matches
URL: https://github.com/apache/hive/pull/647#discussion_r287916346
##########
File path: ql/src/java/org/apache/hadoop/hive/ql/DriverContext.java
##########
@@ -234,4 +240,24 @@ public void apply(FileSinkOperator fsOp) {
}
}
}
+
+ // Create a transaction manager for replication flow. In case user has
provided a different config for REPL load,
+ // this transaction manager will be used to do the transaction operation.
This is specifically useful when REPL load
+ // is called with different hive metastore uri expecting the transaction
information to be stored in the specified
+ // metastore.
+ public synchronized HiveTxnManager getReplTxnManager(HiveConf hiveConf)
throws LockException {
+ if (replTxnManager == null) {
+ replTxnManager =
TxnManagerFactory.getTxnManagerFactory().getTxnManager(hiveConf);
+ LOG.debug("Created a new replTxnManager : " + replTxnManager);
+ }
+ return replTxnManager;
Review comment:
that scenario is not applicable as for one repl command only one txn manager
will be created and we can don't support config change in between.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
Issue Time Tracking
-------------------
Worklog Id: (was: 249088)
Time Spent: 1h 10m (was: 1h)
> Repl load command config is not passed to the txn manager
> ---------------------------------------------------------
>
> Key: HIVE-21766
> URL: https://issues.apache.org/jira/browse/HIVE-21766
> Project: Hive
> Issue Type: Bug
> Affects Versions: 4.0.0
> Reporter: mahesh kumar behera
> Assignee: mahesh kumar behera
> Priority: Major
> Labels: pull-request-available
> Fix For: 4.0.0
>
> Attachments: HIVE-21766.01.patch, HIVE-21766.02.patch,
> HIVE-21766.03.patch
>
> Time Spent: 1h 10m
> Remaining Estimate: 0h
>
> *Cause:*
> REPL LOAD replicates Txn State (writeIds of tables) to the target HMS
> (backend RDBMS). But, in this case, it is still connected to source HMS due
> to configs passed in WITH clause were not stored in HiveTxnManager.
> We pass the config object to the ReplTxnTask objects but HiveTxnManager was
> created by Driver using session config object.
> *Fix:*
> We need to pass it to HiveTxnManager too by creating a txn manager for repl
> txn operations with the config passed by user.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
