On Fri, Nov 18, 2005 at 05:00:06PM +0100, Csaba Nagy wrote: > > Now according to pg_locks, the transaction of the update was holding an > exclusive lock on the table, which I can't explain, as we don't acquire > any table lock in our whole application, and the SQL of the update is On the table, or was it a ROW EXCLUSIVE lock? That will be automatically taken by an UPDATE. > And then, why was I not able to cancel the backend via kill or > pg_cancel_backend ? That's a good question. The return code from kill -2 <pid> (or whatever SIGINT is on your system) might be useful here. Also, attaching to the pid with gdb might help. A -- Andrew Sullivan | ajs@xxxxxxxxxxxxxxx The plural of anecdote is not data. --Roger Brinner
