Hi NetBackupers,

I have had a strange problem and I was hoping someone on this list might
have experienced/fixed this before.  I was getting, intermittently,
Error 54 (* see below) codes on some (not all) of my backups of an
Exchange Server.  I worked real hard to troubleshoot the network, on
both ends (Solaris 9, NetBackup 5.0 Master Server and Windows 2000
client).  Nothing seemed to be wrong.  Frequently, I would have these
errors in my activity log, but when I clicked on the client in the
HostProperties listing in the GUI, it would connect right up.  I would
then be able to manually start a backup.  It was strange.

So, I wrote a script and stuck it in CRON.  The script runs the
bpgetconfig command every minute and records whether a connection was
made or not.  I have found, over 3-4 weeks, that the connection does not
complete sometimes, and this behaviour is very intermittent.  I've
graphed the connection loss to look for patterns and can't find any.
I've implemented the same script to watch other Windows 2000 servers,
and the same thing happens to all of them!!  Usually, the connection
loss only lasts 4-5 mins and then (mysteriosly) comes back.  However,
especially on this Exchange server, sometimes this lasts upwards of an
hour.  I've determined that, when this connection loss overlaps a start
window for a policy, then the policy gets this Error 54 code.  Has
anyone else noticed this sort of thing?

--Errin Larsen


*Error 54 = "The server could not complete the connection to the client.
The accept system or winsock call timed out after 60 seconds."