I think this seems right. I wonder whether we can combine the lock timeout set with the lock acquisition for one fewer back and forth with the database..? Not a big deal right now, if not. Actually, maybe the bigger question is: should we just change the lock timeout at connection time and use that for all lock timeouts?
I believe changing that would affect all implicit/explicit lock requests. Is 5 seconds enough time for such locks? Maybe. We could TIAS but I don't know the possible ramifications
I think this seems right. I wonder whether we can combine the lock timeout set with the lock acquisition for one fewer back and forth with the database..? Not a big deal right now, if not. Actually, maybe the bigger question is: should we just change the lock timeout at connection time and use that for all lock timeouts?