The problem was/is Netbackup sets the media mount timeout to the
CLIENT_READ_TIMEOUT ignoring the global media mount timeout.
If it is unable to assign/reserve tape resources within this time it fails
with error code 6. Good one.

Regards
Peter Marelas

-----Original Message-----
From: Marelas, Peter 
Sent: Wednesday, 12 June 2002 7:39 PM
To: 'veritas-bu AT mailman.eng.auburn DOT edu'
Subject: NBU Oracle Agent for Solaris (RMAN) - sbtbackup function?


Could someone explain what the sbtbackup function does
in relation to NBU Oracle RMAN backups.

sbtbackup is part of the libobk library and is called
for Oracle RMAN backups.

We have found that under moderate system load this function
is taking over 15 minutes to complete even on small databases (< 500MB). If
it takes longer than 15 minutes Netbackup aborts as it only waits for
CLIENT_READ_TIMEOUT seconds (which is 15 minutes by default) for this
function to return. This results in status code 6 failures.

Raising the CLIENT_READ_TIMEOUT alleviates the affect however the problem
does not exhibit itself when backing up to disk versus tape (sbt_tape).

So I would like to know what sbtbackup is actually doing and what we could
do (change the rcv?) to reduce the time this function takes to complete.

Regards
Peter Marelas