We use pooled connections from Cold fusion to grant access to our database from web pages. Sporadically we get one connection which will hold an exclusive table lock, and block all other connections from access to that object. Because the connections are pooled, I can't seem to catch the offending sql, because the inputbuffer for that spid has already accepted new sql from a different cold fusion connection. Has anyone else seen this behavior between SQL server 7.0 and Cold Fusion? so far the only way out of the locking situation is to kill the spid.