-----BEGIN PGP SIGNED MESSAGE-----
Hi Mike,
> > > running: <hdg1><hde1><hdc1>
> > > now!
> > > hdg1's event counter: 00000082
> > > hde1's event counter: 00000080
> > > hdc1's event counter: 00000080
> I followed your instructions, & to be sure loaded all raid
> options into the kernel. Everything seemed to be fine.
> The raid array was accessible. I rebooted twice more.
So far so good.
> I then mounted the degraded array (read only) & tried
> du to check the files. There were a number of IO errors. I
> then rebooted again, & now the two disks are out of sync,
> & once again I have no raid array.
> I cannot help feeling I have made some serious errors. Am I
> still able to recover the situation?
Yes, you should still be able to recover the situaion, but it's going
to be a bit tricky. The one thing that really bothers me is the IO
errors you got when trying to read from the raid array. Can you
verify if you can successfully read the raw devices before going on?
try a dd if=/dev/hde of=/dev/null ..... for the three disks. you
shouldn't go on until raw access actually works. if there's problems
at this step you'll have to get help elsewhere before going on;
hardware funnies aren't my forte.
To recover your raid set you'll have to rewrite the raid superblocks
on the disks. to do so, you need an /etc/raidtab that exactly matches
your current configuration. All settings must match exactly -
sequence, chunk size... .
for /dev/hdg1 change the "raid-disk" line to "failed-disk" in order
to create the ne superblocks in degraded mode - otherwise backbround
resync kicks in rigt after mkraid and you have no chance to retry if
you got something wrong, so make sure you don't skip this!
now, do a mkraid -force; this will change nothing but the raid
superblocks on disk. try mounting the raid device read only, check if
the files are OK. if it doesn't work, recheck that the raidtab really
is OK; if it turns out that one of the disks you're trying to use is
really damaged, you can also try changing the "failed-disk" entry;
from what you've written I assume that the data on all three disks
should still be in sync (you only did read-only access so far,
right?).
One you've got the array up and running again, you can raidhotadd the
final disk to make the array redundant again. at that point, also
change the raidtab failed-disk to raid-disk again to keep things
consistent.
Hope this helps.
Bye, Martin
*******************************************************
Martin Bene, Tech. Manager Web Solutions
KPNQwest Austria GmbH, A-8020 Graz, Nikolaiplatz 4
t +43 (316) 813824, f +43 (316) 813824-26
e [EMAIL PROTECTED], i http://www.kpnqwest.at
*******************************************************
-----BEGIN PGP SIGNATURE-----
Version: PGP Personal Security 7.0.3
iQCVAwUBOuMxDh+NBGYktXFhAQEYnAP/cKoDxxBLPPQfRSvI+OZUg28y01Ep4V+g
G3tIlSuplyDDkFtfquH3UwdRQbDLXPvspnuvOG9QuWdB1NcUCrTN4uFkciQDeh74
W39WvGDmpNrdjrnnRNSnzRzo5iW4Z3hf1On18fH5tn84YvyO/NFpv4eLRHscaq/W
q98axmXakD8=
=4cqc
-----END PGP SIGNATURE-----
-
To unsubscribe from this list: send the line "unsubscribe linux-raid" in
the body of a message to [EMAIL PROTECTED]
