2008/12/3 Michael Lake <[EMAIL PROTECTED]>: > Owen Townend wrote: >> >> I pulled the /dev/md1 from the 'prefered minor' that was in the >> examine output. You can use a different one without hassle. >> The ordering for the command is: >> Usage: mdadm --assemble md-device options-and-component-devices... > > So it assembles physical devices onto the virtual /dev/md0 > >> The '--run' is because with a missing device it would usually only >> assemble the array and leave it stopped which is the safest default. > > OK > >> Try >> # /sbin/mdadm --assemble /dev/md0 --run -u >> b6035660:c11a8a2e:7026aea2:99d23bfc /dev/sdc5 > > No recognisable superblock on /dev/sdc5 > dev/sdc5 has wrong uuid > No devices found for /dev/md0 > > BUT! I tried it with sdc1 using it's uuid and it works! > I message came up that /dev/md0 has been started with 1 drive out of 2 > Then mount /dev/md0 /mnt worked and I can see the files and directories. > > I can now see the old /etc/fstab and it tells me that it was a reiserfs > system. > > What I don't follow is that now If I run: > /sbin/mdadm --examine /dev/sdc5 > > "No md superblock detected on /dev/sdc5" where as before it said > /dev/sdc5: > Magic : a92b4efc > etc.... > > I have some playing to do to work it out but am now much relived I can see > data. > > Mike
Hey, I know I put in the disclaimer that I hadn't checked the syntax, but I still feel bad that I now realise that the initial command may have tried to use /dev/sdc5 as the node for the resultant raid array. This may be the culprit for the missing superblock. You are in a good position though having both the original --examine data (if you wish to recreate the superblock) and a second copy of the data (the other disk from the mirror). Hope it works out ok... cheers, Owen. N.B. Read the man pages, check the syntax, only then then run commands as root on data you're trying to recover... -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
