On Wed, Jun 29, 2005 at 04:34:19PM +1000, Rebecca Dridan wrote:
> I hope someone can help me with an odd amanda problem. I've searched the
> net and archives, but haven't seen anyone else with this problem yet.
>
> Last week I had a fully working Amanda setup on a BSD server, with a
> daily and a weekly set of tapes. This week, when I put the next tape in,
> and run amcheck daily, I get:
>
> ERROR: /dev/nsa0: not an amanda tape
> (expecting tape Daily03 or a new tape)
>
> That is happening with _all_ tapes I've checked, daily or weekly. The
> two sets are stored in different places and I really can't accept the
> fact that they all died at the same time. The only thing that happened
> between amanda working and not working is that the server was rebooted.
>
> I labelled a new tape, and flushed a backup to it, and that tape appears
> to work, but that doesn't solve my problem of getting to my old backups,
> or of whether I have to label a new set of tapes every time the server
> is rebooted. Does anyone have any ideas what to look for? Even dd is not
> finding anything on the old tapes, making me scared to check too many of
> them, in case the act of checking them is in some way wiping them....
>
> Hope someone can give me a clue.
You've eliminated most of my first guesses, hardware, environment, ...
I'm left looking at that reboot, and possibly a power cycle affecting
the settings on the tape drive.
Perhaps the tape drive has a default block size which had been
changed when the old tapes were used?
You say dd did not find anything? Can you be more specific about
your testing with dd? It should be safe as long as you don't
use "of=your_tape" or "> your_tape".
dd sometimes has trouble reading tapes when its own block size is
smaller than the blocks on the tape. Something like "out of memory"
or "insufficient space" is the message. Try some block sizes other
than the default (0.5KB on my dd, much too small). Amanda probably
does not go lower than 32KB. But your tape may have been set to
an even larger size. The dd options are ibs= or bs=. I don't think
setting it too large is a problem. Perhaps try bs=1024K right away.
--
Jon H. LaBadie jon AT jgcomp DOT com
JG Computing
4455 Province Line Road (609) 252-0159
Princeton, NJ 08540-4322 (609) 683-7220 (fax)
|
