rdblue commented on pull request #2754:
URL: https://github.com/apache/iceberg/pull/2754#issuecomment-872632160


   @jshmchenxi, it sounds like the problem might be that the UGI in your 
environment changes based on the incoming request, but the thread pool that 
handles planning keeps its initial user information?
   
   I'm suspicious that the solution is to add Hadoop-specific UGI code 
everywhere that Iceberg uses a thread pool. This solution doesn't seem 
sustainable to me. One easy solution is to disable the thread pool in your 
environment by setting `iceberg.scan.plan-in-worker-pool=false` in Java's 
system properties. If you want to use a thread pool, we could update the 
`TableScan` interface and Spark to manage thread pools at a session level.


-- 
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]



---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to