okumin commented on code in PR #395:
URL: https://github.com/apache/tez/pull/395#discussion_r1975090250
##########
tez-mapreduce/src/main/java/org/apache/tez/mapreduce/client/YARNRunner.java:
##########
@@ -623,6 +623,7 @@ public JobStatus submitJob(JobID jobId, String
jobSubmitDir, Credentials ts)
try {
dagAMConf.set(TezConfiguration.TEZ_AM_STAGING_DIR,
jobSubmitDir);
+ dagAMConf.setBoolean(TezConfiguration.TEZ_AM_STAGING_BASE_DIR_CLEANUP,
true);
Review Comment:
YARNRunner is a glue code between a MapReduce Job and Tez, implementing
`ClientProtocol`. So, the client code of YARNRunner is Apache Hadoop.
`ClientProtocol` doesn't have an API to declare that a specific job has been
completed. If we resolve this issue on the client side, we have to add new APIs
to Apache Hadoop. That's why I added a new param and handled the issue on
Apache Tez side.
I'm not confident that this approach is the best. I'd appreciate it if
someone could give me a better idea.
--
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.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
