Veritas-bu

[Veritas-bu] How do you handle status code 1?

2004-06-11 15:30:25
Subject: [Veritas-bu] How do you handle status code 1?
From: WTS AT Maine DOT edu (Wayne T Smith)
Date: Fri, 11 Jun 2004 15:30:25 -0400
> 2. Does anyone have a good script to show which files are being missed
> on a given server? I know I can get the info in bperror and the NBU
> reports, but the format is awful.
> 
> 3. What does your company do to handle status code 1's? I'm looking for
> both the technical aspect as well as the processes. 

I've tried to teach people that status code 1's increase risk of not 
having important information backed up and available for restore when 
needed.

So it's important to understand why one gets a status code 1 and avoid 
it if possible.

To those ends, I extract all messages for each client (for the past 24 
and 96 hours) and put it on a web site along with other status info. 
The bperror "-U" output is pretty readable ... so I use this (pardon the 
Rexx, please):

    address SYSTEM ,
    bperror "-columns" fold_col ,
            "-U -all -hoursago 24 -client" client.iclient ,
            ">" cpath||client.iclient"/all_log24.txt" with error stem e.

Once people know the files that are missed, they can evaluate each 
problem and workaround or bypass.  For example, file "parent.lock" 
cannot be read by NetBackup sometimes if the user leaves Netscape 
running.  It's an uninteresting file, so I or the user excludes 
"parent.lock" from future backup consideration.  If the single file 
failures are database or (Outlook) ".pst" files, then something more 
needs to be done ... such as not leaving Outlook running overnight.

If people can change their status code 1 results to 0, then a "1" won't 
be background static/noise, but something to notice.  Unfortunately, 
this is often an unattainable goal for Windows systems and NetBackup.

Hope this helps!

cheers, wayne

<Prev in Thread] Current Thread [Next in Thread>