(OT) cool data recovery story from a bad drive (LVM)
Recently one of my 2 drives has gone bad.
Not that big a deal but i didn't look forward to reinstalling
the whole shabang.
I use LVM and both drives are used in the same volume group so i
wasn't sure that it was possible to save the data and keep the
machine running without having to reinstall it.
Well, i managed to keep the server running while saving the data
that resides on the bad disk.
Mind that only the /usr partition used PE's from the 2nd disk
(the one that started failing).
First i made a partition to copy all the data over from the
old /usr. A lot of data was unrecoverable because of errors on
that disk. I then adjusted fstab to point to the new partition
to see if that would work. It did. Next i removed the bad disk
from the volume group via lvreduce.
I still have some PE's left to make a new /usr partition in the
LVM partition. After that i copied the data over from the
/usr partition to this newly created /dev/main/lv_usr (main being
the volumegroup), again adjusted the fstab to reflect the change
and rebooted.
All services (fetchmail,exim4,spamassassin,clamav,procmail,apache
,courier,squirrelmail) kept on working just fine. To rebuild my
/usr partition i started to do "apt-get --reinstall install <package>"
for most packages that where installed. Most of the usr partition
have now been rebuild and everything is working except my ddt-client.
So i'm happy i didn't have to reinstall and it was a learning
experience. I'll be closely monitoring the system to see if everything
is stable but so far so good.
Regards,
Benedict
Reply to: