At 11:40 AM 1/26/01 -0600, David A. Chapa wrote:
>C-
>
> >>So what you are saying is that I could do something like this:
>
>Yes, exactly.  And you can get the pool number by using vmpool -listall if
>you'd like.

That's actually what I do.  I just didn't put that in my example code. ;)

> >>So would that change the volume's status to the same "0x01" status, like
>doing it the
> >>other way?
>
>I'm not sure what you mean here, I know when you assign tape to be dedicated
>to the DB Backup process, netbackup will assign those tapes, even though
>they have never (potentially) been written to, it assigns them for exclusive
>use to this utility (bpbackupdb).

So it's the bpsyncinfo command, or the 'database backup setup' utility that 
makes this change?  If that's the case, then the database backup tapes 
under YOUR setup would never get that weird "0x1" status that "normal" 
database backup tapes get.  (You can see it with a vmquery -m <volume>.)

>The problem when we do these backups outside of this internal process is
>that WE need to keep track of these tapes and which have been used and which
>one should be next, etc.  Which is why a separate Volpool is a good idea.
>That is what I would use the DBBACKUP_CALLED text file for because each time
>we do a dbbackupdb, part of the information that is written to this file is
>the media id that was used for the process.  Therefore, if your client wants
>to have say 5 db backup tapes in a rotation, then I would use that text file
>and perform my tests to determine the next tape in the sequence to be used.
>(does that make sense?)

My only problem with that is that you're not controlling that file.  If 
someone else grabs the GUI, or manually initiates a database backup to 
another tape, your setup would get wonkied.  That's why I'm tracking myself 
which tapes _MY_ script is backing up to.  (You can see that in the latter 
half of the script, if you look at the full version I forwarded.)

>By the way you are correct about the -f mountpoint_file switch with tpreq.

Cool!

>And incidentally, the -bx switch for vmquery will give you an all columns
>display, much like you would see in the motif interface (xvmadm).

Hmm... Wouldn't it be nice if that was documented? ;)

>Hope this helps.

Sho' does!