Hi!
Trying to kill the keyboard, [EMAIL PROTECTED] produced:
> I didn't clean the head because it's a new hardware and as I have written
> it works under Win95. I retensioned both tapes several times.
After I did clean the head of my tape device the first time
(when it was barely used) I was astonished how much dirt was
there. I assume that not only the then-new tapes left above
average residue but that the head was not very clean coming from
the factory :-)
> Logfile shows:
> Jan 14 19:29:39 TeRrOrItE kernel: [218] 0 ftape-write.c
> (ftape_write_segment_R6ec54595) - write error, retry 1 (5459).
Soft errors. Writing the data failed once and is retried.
The retry works (else there would be another warning with retry 2.
There can be up to seven retries before ftape gives up. As long
as it works in the end, no need to worry.
> [cut more errors]
> Jan 14 21:41:38 TeRrOrItE kernel: [252] 0 ftape-write.c
> (write_segment) - warning: 1 hard error(s) in written segment.
> Jan 14 21:41:38 TeRrOrItE kernel: [253] 0 ftape-write.c
> (write_segment) - hard error in segment 47464.
> Jan 14 21:41:40 TeRrOrItE kernel: [254] 0 ftape-write.c
> (ftape_write_segment_R6ec54595) - write_segment failed, error: -5.
> Jan 14 21:41:40 TeRrOrItE kernel: [255] 0 zftape-write.c
> (zftape_write_segment) - Hard error writing segment, trying to recover.
> Jan 14 21:41:40 TeRrOrItE kernel: [000] 0 ftape-bsm.c
> (ftape_put_bad_sector_entry) - putting map 0x00100000 at 02866352, segment
> 47464.
> Jan 14 21:41:40 TeRrOrItE kernel: [001] 0 ftape-write.c
> (write_segment) - TAIL->status == waiting.
> Jan 14 21:41:40 TeRrOrItE kernel: [002] 0 ftape-write.c
> (write_segment) - HEAD: 01d48698, TAIL: 01d48698.
> Jan 14 21:41:40 TeRrOrItE kernel: [003] 0 ftape-write.c
> (write_segment) - runner_status: 0.
> Jan 14 21:41:41 TeRrOrItE kernel: [004] 0 ftape-write.c
> (ftape_write_segment_R6ec54595) - write error, retry 1 (47464).
> Jan 14 21:41:41 TeRrOrItE kernel: [005] 0 zftape-write.c
> (empty_hard_error_list) - Trying to handle hard error.
There is a hard error on the tape (a spot which should have been
marked bad during format, but was slipped or develloped later).
There is also a soft error on the same sector (see above).
Both errors are handled, it seems, the backup should be 100% OK.
However, you should *always* compare your backup, just to be sure.
[statistics]
> Jan 14 21:43:45 TeRrOrItE kernel: no_data_errors : 35
Data was a bit too late in arriving at the tape, I think.
> Jan 14 21:43:45 TeRrOrItE kernel: retries : 34.
Soft errors produce retries.
> Jan 14 21:43:45 TeRrOrItE kernel: [008] 0 ftape-ctl.c
> (ftape_print_history) - ecc statistics:
> Jan 14 21:43:45 TeRrOrItE kernel: crc_errors : 0
> Jan 14 21:43:45 TeRrOrItE kernel: crc_failures : 0
> Jan 14 21:43:45 TeRrOrItE kernel: ecc_failures : 0
When reading from the tape, ecc_failures should be zero. They
would be sectors where there was an error which could not be
corrected by reading the ecc bits.
> Jan 14 21:43:45 TeRrOrItE kernel: sectors corrected: 0.
> Jan 14 21:43:45 TeRrOrItE kernel: [009] 0 ftape-ctl.c
> (ftape_print_history) - Warning: 1 media defects!.
That is the hard error from above. If many of those appear,
consider the tape bad. AFAIK the Pro cannot format tapes, which
is bad luck ... OTOH, an 1.6/3.2GB tape takes easily 8-10 hours
to format.
> Jan 14 21:43:45 TeRrOrItE kernel: [010] 0 ftape-ctl.c
> (ftape_print_history) - tape motion statistics:
> Jan 14 21:43:45 TeRrOrItE kernel: repositions : 163.
Each time the motor starts (I think) is a reposition, whether it
is because the tape is at the end of it's length and we change
track, or an retry or a no data error. You can try to lower the
no data errors by using the program 'buffer' between your backup
program and /dev/TAPE, e.g.
afio -o ... - | buffer -m 5m -p 80 -u 100 -s 10k -S 100k > /dev/tape
(5 mb buffer, start writing after 80% filled, sleep 100 ms after
each write (dunno if that's effective), blocksize 10k, print
amount each 100k to screen)
The reason your tape seems to work in windows is that windows
hides these diagnostic messages.
-Wolfgang
--
PGP 2 welcome: Mail me, subject "send PGP-key".
Unsolicited Bulk E-Mails: *You* pay for ads you never wanted.
How to dominate the Internet/WWW/etc? Destroy the protocols! See:
http://www.opensource.org/halloween.html