-
Notifications
You must be signed in to change notification settings - Fork 2.2k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Token expiration issued when using WORKER_POOL with Spark Thrift Server #2753
Comments
Is this a problem of using wrong UGI, which caused access control problem, or just the token is expired? PS. I'm not familiar with Kyuubi. Does kyuubi support using kerberos to renew tgt periodically? |
It's a problem using wrong UGI. The UGI is correct outside the WORKER_POOL.
Obviously the UGI that WORKER_POOL is using is not correct. |
This issue has been automatically marked as stale because it has been open for 180 days with no activity. It will be closed in next 14 days if no further activity occurs. To permanently prevent this issue from being considered stale, add the label 'not-stale', but commenting on the issue is preferred when possible. |
This issue has been closed because it has not received any activity in the last 14 days since being marked as 'stale' |
We are using a Spark Thrift Server (Kyuubi) to provide adhoc query service for Iceberg.
The global WORKER_POOL is enabled by default.
We found that after several days of running, some queries like
update test_table set data = 'abcd' where id = 1
start to throw such exception:The reason is that WORKER_POOL is a singleton thread pool, initialized when service start. The first time a thread in WORKER_POOL is accessed, the ugi or credentials (which I'm not very clear about) of the context to the first query are left in the thread. The second time another query is issued and access the same thread, the credentials of that thread are still from the first query, causing this problem.
We should use ugi.doAs() when using WORKER_POOL if tasks include accessing the cluster(eg. HDFS)
The text was updated successfully, but these errors were encountered: