[plug] Weird "tape full" error

[Andrew Furey]

Hi all,

We have a client with a Debian potato box, which has a
FutureDomain SCSI tape drive in (model TMC-18C30),
using /dev/st0 with the st and fdomain modules.

We originally set it up last November to do a daily
back up of the client's important directory tree
(shared drives). It was working fine back then, but
since February 22, every day the command

cd /home/public; tar -cvvzf /dev/st0 *

has come back with

tar (grandchild): /dev/st0: Cannot write: No space
left on device
tar (grandchild): Error is not recoverable: exiting
now

Obviously it thinks it has run out of space, so here's
some stats:

* Tapes we are using are 2Gb, hence with a theoretical
compressed value of 4Gb.
* Total uncompressed size of the source tree is
3.37Gb, and has not grown significantly since
November.
* According to a backup-to-file test, this compresses
down to only 1.87Gb, which should definitely fit on
the tape.
* Doing a restore of the failed backup produces a tree
of almost exactly 3Gb, but given the compression I'm
not sure how significant that is.

Also, looking at the tapes they use, it would _appear_
that they are only used up to about half-way, but I
can't confirm this for sure because it seems the drive
itself insists on rewinding them before ejecting, even
when using /dev/nst0...

Anyone seen anything like this before?

TIA,
Andrew

=====
"One Architecture, One OS" also translates as "One Egg, One Basket".

http://www.sold.com.au - SOLD.com.au Auctions
- 1,000s of Bargains!