-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Hello everyone, again! It looks like my problems with my new OnStream
DI-30 drive are continuing. Currently I can create and verify tapes
if I create them using a single tar command; however, tapes written by
amanda (using tar) cannot be verified. I believe this problem comes
from some problem with the driver tape positioning logic.
In general, I find that "mt status" has serious problems. It
generally gives the right answer if the tape is fully rewound;
however, if I fast-forward to another file mark (mt fsf), it often
reports a -1 for the block count. It always seems to report 0 for the
file count.
Even more disturbing, though, is that the block numbers can change
spontaneously! After doing a "mt fsf" on a tape that has several
small (4 meg) tar archives on it (which were written consecutively),
multiple "mt status" calls report increasingly higher block counts.
This is with no other tape access calls being made! Each call to "mt
status" also causes the drive to run for a second or two.
Multiple "mt status" commands run in quick succession when the tape is
not completely rewound can cause hundreds of these messages to be
logged:
Jun 3 11:04:27 lydia kernel: ide-tape: ht0: skipping frame,
wrt_pass_cntr 12 (expected 15)(logical_blk_num 276367)
Jun 3 11:04:27 lydia kernel: ide-tape: ht0: skipping frame,
wrt_pass_cntr 12 (expected 15)(logical_blk_num 276367)
Jun 3 11:04:27 lydia kernel: ide-tape: ht0: skipping frame,
wrt_pass_cntr 12 (expected 15)(logical_blk_num 276368)
[many, many messages...]
Jun 3 11:05:28 lydia kernel: ide-tape: ht0: skipping frame,
wrt_pass_cntr 12 (expected 15)(logical_blk_num 277367)
Jun 3 11:05:28 lydia kernel: ide-tape: ht0: skipping frame,
wrt_pass_cntr 12 (expected 15)(logical_blk_num 277367)
Jun 3 11:05:28 lydia kernel: ide-tape: ht0: couldn't find logical
block -1, aborting
After trying to verify a tape written by amanda (which backed up 7
partitions on two machines, about 9G total), I got the following errors:
Jun 2 23:44:35 lydia kernel: ide-tape: ht0: skipping frame,
logical_blk_num 317810 (expected 0)
Jun 2 23:44:35 lydia kernel: ide-tape: ht0: couldn't find logical
block 0, aborting (block 317810 found)
Jun 2 23:44:35 lydia kernel: ide-tape: ht0: unrecovered read error
on logical block number 0, skipping
I actually got hundreds more of these messages, reporting problems
finding logical blocks in the 0-14 range, and instead finding 300,000+
logical blocks. To me, it almost sounds like the tape got rewound at
some point in the run, but the driver thought it was writing at the
end of the tape, so it overwrote the beginning of the tape.
So, now I'm wondering whether my problems are because of bad tapes,
some driver problem, or maybe even a bad drive? I am currently using
a Debian 2.2 system, with a 2.2.15 kernel patched with ide-20000509
and the recently posted fix for the "block 3015" problem (which I was
also experiencing).
Anyone seen similar behavior?
--Anil
- --
Anil Somayaji ([EMAIL PROTECTED])
http://www.cs.unm.edu/~soma
+1 505 872 3150
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.1 (GNU/Linux)
iEYEARECAAYFAjk5O5oACgkQXOpXEmNZ3SeUfgCfURBS4jlYSu/2JlURqemPuzQL
w94AoIMsnFKSU2o0uqhZfR1P/gDjfett
=ifOd
-----END PGP SIGNATURE-----
