> From: "David A. Chapa" <david AT datastaff DOT com>
> To: "'Louis-Luc Le Guerrier'" <leguerri AT canr.hydro.qc DOT ca>,
<veritas-bu AT mailman.eng.auburn DOT edu>
> Subject: RE: [Veritas-bu] What's the rule for tape-picking?
> Date: Thu, 22 May 2003 12:48:41 -0500
> MIME-Version: 1.0
> Content-Transfer-Encoding: 7bit
> X-Priority: 3 (Normal)
> X-MSMail-Priority: Normal
> Importance: Normal
> X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1165
>
> Have you gotten an answer on this yet?
>
> Well from the Media server perspective, when it is handed a
job from the
> Master, it takes particular notice of a couple of things in
order to
> make the right media selection.
>
> For Example:
> 1. Volume Pool
> 2. Retention
> 3. Density
> 4. Media State (full, frozen, suspended, etc.)
>
> If it finds that one of the media in its DB satisfies its
requirement
> request (and the tape is not in use), then it will mount one
of the tape
> in its MEDIADB, if not it will make a request to the
volHOST (typically
> the master server) for a tape that meets its requirements
>
> For Example:
> 1.Volume Pool
> 2.Density
> 3.Condition (number of mounts, etc.)
>
> It doesn't select the media numerically based on the ID, but
numerically
> based on the number of mounts. The one with fewer mounts
should be
> selected first. If you check the number of mounts on tape
000191, you
> may find that it was less than 000190.
>
The problem is all the tapes are in same volume pool, same
density, and yes I did know about the lowest # of mounts
selected first. I set #mounts to the ones I want to 0 first.
The thing is all tapes are likely to be chosen, but I can't
determine on what the system bases on to select a particular
tape rather than another. Last week-end, it did something
strange, too.
I still have the 000188 pending, not chosen. It has picked
000191. When this one got full, it didn't pick 000188 nor
000192 which were both active, and I desired to fill them up
first. Rather it took an empty one, 000193.
Even worse! When 000193 got full, it jumped to 000196, and
left behind 000194 and 000195.
So another question, why did it pick 000196, and not 000194
as one would expect, or better, finish filling up 000188 or
000192?
I set all the # of mounts to 0.
Furthermore, I set them up in the filling order I want them
in the robot slots, meaning I put 000188 in slot 1, 000191 in
slot 2, 000192 in slot 3, 000193 in slot 4, 000194 in slot 5,
000195 in slot 6, 000196 in slot 7 and 000197 in slot 8.
So I don't grasp the logic why 000196 got selected rather than
another one.
Thanks for any enlighting.
Louis-Luc Le Guerrier
> Not knowing your environment it is difficult to comment on
why it didn't
> select your active media for your backup jobs, but perhaps
some of this
> may help.
>
> David
>
> PS. This selection flow is a very simplistic view of a
complex process,
> more goes into it than I am describing.
>
> >-----Original Message-----
> >From: veritas-bu-admin AT mailman.eng.auburn DOT edu
> >[mailto:veritas-bu-admin AT mailman.eng.auburn DOT edu] On Behalf
Of
> >Louis-Luc Le Guerrier
> >Sent: Monday, May 12, 2003 10:36 AM
> >To: veritas-bu AT mailman.eng.auburn DOT edu
> >Subject: [Veritas-bu] What's the rule for tape-picking?
> >
> >
> >Hello,
> >I'm using NB 3.4, and I have a volume pool with different
> >tapes available. Let's give an example:
> >
> >Tape Bytes used Status
> >000188 5236246 Active
> >000189 2634674 Active
> >000190 - Available
> >000191 - Available
> >
> >All these tapes are in the same volume pool, and likely to
> >be picked to put the data of the next backup job.
> >I have 2 drives, so on the next backup session when jobs
> >start, I want to use 000188 and 000189 in each drive, until
> >they're full, before picking another tape.
> >
> >Usually Volume Manager does it right, but last time in this
> >example, it picked 000189, and 000191, leaving 000188
behind.
> >Furthermore, I'd expect 000190 to be used before 000191,
> >because of numerical order.
> >
> >Does anyone know the tape picking rule? Or is there a way
to
> >change it? Ideally, I'd like to make the ACTIVE tapes
picked
> >first, then the lower-numbered labelled tapes among the
AVAILABLE ones.
> >
> >Thank you for any help.
> >Louis-Luc
> >
> >P.S.: I'll soon upgrade to NB 4.5, so is there any
difference?
> >
> >_______________________________________________
> >Veritas-bu maillist - Veritas-bu AT mailman.eng.auburn DOT edu
> >http://mailman.eng.auburn.edu/mailman/listinfo/>veritas-bu
> >
>
|
