To preface this, I am aware of the optimistic locking scheme used between nodes in the cluster, and that what is being reported back as a deadlock, is not truely a deadlock, but a notification of an intersect on row updates between nodes. (I think I have stated that correctly). But having said that I am seeing deadlocks reported as errors in our app more often than I would expect, and I’m seeing it reported on select statements. Which doesn’t make sense to me. The select and active record error message are below. The tables in the query are very rarely updated in the app, so I doubt we’d see multiple updaters on either table. The only thing I can think of, is that the statement is part of a transaction, and other sql in the transaction could cause a multi update error. But if that were the case, why flag the select as deadlocked ?
[IMG]https://mail.google.com/mail/u/0/images/cleardot.gif[/IMG]