pvary commented on PR #5036: URL: https://github.com/apache/iceberg/pull/5036#issuecomment-1157393827
> > After fixing this issue, you still see long alter table HMS calls? I am trying to understand the situation better, and possible other places where we have to fix too > > Yea, though likely it won't be as frequent anymore. I think the point I am trying to make is that we can not guarantee that Iceberg commit would be able finish within hive.txn.timeout. You and I both listed a few reasons on why that can happen. I think heartbeats on locks is a must have, or we should fail the commit/ rollback the commit if unlock fails with lock not found. I prefer the former as it would really be a waste of compute to simply fail or rollback. You have convinced me here, I just trying to collect the things that we have to fix in Hive to make this work as fast as possible 😄 -- 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]
