[
https://issues.apache.org/jira/browse/HIVE-24695?focusedWorklogId=543450&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-543450
]
ASF GitHub Bot logged work on HIVE-24695:
-----------------------------------------
Author: ASF GitHub Bot
Created on: 28/Jan/21 08:40
Start Date: 28/Jan/21 08:40
Worklog Time Spent: 10m
Work Description: rbalamohan opened a new pull request #1921:
URL: https://github.com/apache/hive/pull/1921
https://issues.apache.org/jira/browse/HIVE-24695
### What changes were proposed in this pull request?
There are cases when TezSessionState would not be able to start. (e.g
resource constraints on YARN queues).
However by this time, session directories & certain resources are localized.
(e.g, hive-exec jars are stored in
hdfs:///tmp/hive/hive/_tez_session_dir//hive-exec.jar). When tezClient is not
started, it does not clear up the resources. This leaks ~100MB data in HDFS per
failure.
### Why are the changes needed?
Need to clear up resources in session directory, when errors are encountered
in tez session init phase.
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
Small scale cluster
----------------------------------------------------------------
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: 543450)
Remaining Estimate: 0h
Time Spent: 10m
> Clean up session resources, if TezSession is unable to start
> ------------------------------------------------------------
>
> Key: HIVE-24695
> URL: https://issues.apache.org/jira/browse/HIVE-24695
> Project: Hive
> Issue Type: Improvement
> Reporter: Rajesh Balamohan
> Priority: Major
> Time Spent: 10m
> Remaining Estimate: 0h
>
> There are cases when TezSessionState would not be able to start. (e.g
> resource constraints on YARN queues).
> However by this time, session directories & certain resources are localized.
> (e.g, hive-exec jars are stored in
> hdfs:///tmp/hive/hive/_tez_session_dir/*/hive-exec*.jar). When tezClient is
> not started, it does not clear up the resources. This leaks ~100MB data in
> HDFS per failure.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
