On Wed, Mar 08, 2000 at 08:00:55AM -0500, Henfling, Timothy wrote:
> I'm experiencing some problems with the SQL client in doing backups.  Most
> of the time the backups run ok, however on occasion I will receive a 54
> status error failed to connect and the next couple of backups (transaction
> logs) in the same night will fail with the same status error.  Then it will
> backup the next couple transaction logs later that evening.  The OS always
> backups never time outs.  Any suggestions to correct this.  Do I need to
> increase the time out for connect?  I'm using ver 3.1.5 moving to 3.2.

I know this is a long delayed response, but I never saw a followup to
this question.  We are experiencing a similar problem with Informix log
backups under Solaris.  Once or twice a day, a log backup will fail with
error 54, but filesystem backups never have a problem.  We are using
NetBackup 3.2.

I recently noticed something interesting.  One of our media servers
(not the one doing the log backups) was down for a while for
maintenance.  While it was down, about one third of our log backups
failed with error 54.  As soon as the media server came back up, log
backups went back to their old pattern of failing only once or twice a
day.  I'm not sure why having one media server down would cause connects
to a different media server to time out.  Perhaps changing
SLAVE_CONNECT_TIMEOUT would help, but I haven't had a chance to try
this.

--
Deron Johnson
djohnson AT amgen DOT com