Greetings to all !
I had asked on this list, and the answer came from
John J. Francini <francini_at_progress.com>
(many thanks !) whose text I insert at my original questions:
> >1) Should 60m and 90m tapes be exchangeable between TLZ06 and TLZ07 ?
> > Only uncompressed, or also compressed ?
>
> Yes, in either format.
>
> >2) If yes: Are these read errors an indication of wear,
> > of misalignment, or are there any other guesses ?
>
> Likely to be misalignment, especially in the older of the two drives.
> One way to check for sure is to rewrite the tape on the TLZ06 with
> compression turned OFF (use /dev/rmt0a instead of /dev/rmt0h when
> writing the tape).
>
> >3) Should the error log (which one ?) contain information for more
> > exact diagnostics, and would you consider it worth the effort
> > to analyze that - or is DAT technology simply to be discarded ?
>
> You could use uerf or dia (if installed) to look at the system binary
> error log for errors reported on the tape drives. I don't
> necessarily consider DAT technology as something to be discarded; I'd
> consider getting the drive replaced if the system is under contract...
In the meantime, I have found the 'tapex' command and run more
tests. The situation is now:
a) My backup set consists of 7 tapes (60m + 90m). Some tapes
showed write errors, I used others to replace them,
so I ended up with 7 which seemed to write o.k.
Of these 7, 1 showed read errors on the original TLZ06 drive
used to write it, this one plus 2 others also failed to read
on the TLZ07
(reading tested by 'restore -t', 'dd', or 'tapex -m -w').
b) To repeat a single file system backup, I did this sequence:
# insert another tape in the TLZ06
tapex -E # full test, completed o.k.
dump -0 -u -b 64 -f /dev/rmt0h /u2 # protocol o.k.
mt offline
# slide switch to write-protect, re-insert tape in same drive
/usr/field/tapex -m -w # read tape, count files + records
and received a read error !
It seems to me this drive is severely defective, but unluckily
we have no support contract for this system - I leave the
decision to management whether to pay for repair/exchange
or immediately replace it by a DLT.
c) The man page for 'tapex' memntions an upper record length limit
of 64.512 which is less than the 65.536 used by 'dump .. -b 64',
should I reduce the dump block size ?
Regards, Joerg Bruehe
--
Joerg Bruehe, SQL Datenbanksysteme GmbH, Berlin, Germany
(speaking only for himself)
mailto: joerg_at_sql.de
Received on Wed Feb 23 2000 - 09:33:25 NZDT