I have a maintenance script implemented as a SP running on approx. 50 SQL servers - one of
the things it does is run a backup to disk 'with init' of all databases. Most of the time
the procedure runs fine, but sometimes it ends up 'hanging' on one of the Dbs. The backup
of that db is only about 10Mbytes, and normally completes very quickly, but for some
reason the process sometimes just stays there indefinitely (several days). Attempts to
'kill' the spid fail. Only way to kill it is to recycle the service. There are no enties
in 'errorlog' or NT event log to hint at anything wrong, and everything else appears to
work normally.
My own analysis of 'sysprocesses' and 'sp_lock' so far fail to enlighten me - please help!
(I'm including an extract of 'sysprocess' and 'sp_lock' as an attachment to a separate
post).
--
Erik Oxaal
Conceptos a.s