Re: RAID failure
On Tue, Apr 20, 2004 at 02:26:53PM -0700, Alvin Oga wrote:
> hi ya antony
> On Tue, 20 Apr 2004, Antony Gelberg wrote:
> > The next message was hda 0x51 DriveReady SeekComplete, followed by the
> > actual kernel panic in the RAID code. I disconnected hda, leaving hdc
> > connected. I then got a successful boot, which seems to indicate that
> good that you can boot off hdc ...
> > hda has a hardware fault. But I seem to be missing files from the last
> > five days. How can this be? The system was working fine for most of
> > the last five days, and I am confused as to why these missing files
> > weren't raided to hdc.
> if you were NOT monitoring for drive failures, you probably had a dead
> hda for the past 5 days .. and i finally gave up when you noticed
> - you will not notice that hda is dead as hdc will continue
> to work in degraded mode ... but no copy of it on the other disk
> i guess if hda has file1, file2, file3, file4, file5
> and if hdc has only file1, file2
> when you plug in a new replacment disk ... raid1 will
> "properly resync" things for you ... but...
> when does new raid1 erase file3, file4, file5 from hda
> or when does new hdc raid1 add file3,4,5 to hdc
> both disks is supposed to be identical in data
Nice little mind twister there. What I would do is (if possible)
mount the /dev/hda partitions and copy the important stuff
you need. Then copy the data off your RAID machine completely before
you add a new /dev/hda disk.
I have the same problem on my practice RAID machine. One drive
keeps failing, but every couple of boots I am able to add
it back into the array. So you might be able to mount /dev/hda
once again and look around for your missing data and get it onto
another disk or machine.