[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Questions about software RAID level 1 (how long should it take to sync, etc?)

On 4/4/06, Jim Barber <jim.barber@ddihealth.com> wrote:
> Mike Garey wrote:
> > can anybody tell me why /dev/md0 is no longer available on /dev/hdc1,
> > and if possible how to recreate it?
> Back when I was playing around with trying to get a RAID 1 boot device working again after I had failures on both drives I found
> myself having to go back to square one and trying again.
> It was my fault, I had been getting errors on both drives once every month or two and I'd just tell the failed disk to rebuild again
> and all would be good for another month or two (it was a home system so nothing critical).
> Anyway, eventually I got errors on both drives at the same time, and I could no longer rebuild one disk because it couldn't read the
> disk that was still part of the array.
> Also the corruption on the primary drive was such that the system wouldn't reboot.
> It got disk seek errors during the init process and a kernel panic followed.
> I know the disk that got marked as failed was actually the better of the two disks (it only had errors once every 6 months or so ;)).
> So I want to boot off that, but Linux refused to use it when booting from a Rescue CD because it had been marked as a failed disk in
> an array.
> In the end I had to zero the super blocks on the disk before I could get the rescue disk to mount it as the root file system.
> I did it by using the --zero-superblock option of the mdadm tool.
> Once that was done, it was no longer a RAID disk, and I could just mount it as /dev/sda1, and I was on my way to copying the data
> over to my new HDD ready for setting up the RAID again.
> So perhaps that will be useful to you to get back to a point where you can diagnose your slow transfer option and then start the
> rebuild process again.

Thanks for the suggestion Jim, although I really would like to figure
out how I can salvage /dev/hdc1, since if I can't manage to get it to
work now, what chance will I have when a disk _really_ fails?  I have
a feeling now that it may have something to do with the initrd image
created by mkinitramfs, and the script used to start md, perhaps it's
not creating /dev/md0 properly..  In any case, I may have to end up
zeroing the superblocks, but I'd like to exhaust every other avenue
first.  Thanks again for the info,


Reply to: