Google Groepen ondersteunt geen nieuwe Usenet-berichten of -abonnementen meer. Historische content blijft zichtbaar.

Clone or copy failing hard Disk

32 weergaven
Naar het eerste ongelezen bericht

Gordon Darling

ongelezen,
30 jan 2008, 09:25:4130-01-2008
aan
I have a dual core 64 bit Athlon based box running Suse 10.3 with a
failing hard disk. It'll run for 4 to 8 hours then freeze. Reboot and the
disk is missing from the BIOS. Wait 30 minutes for it to cool down and
the disk reappears and then I have to fsck the disk. The box is still
under rtb warranty but the supplier is sending me a replacement disk on a
swap basis.

I've managed to back up the important stuff to DVD (/home /etc and so on)
but the ideal solution would be to try and clone the disk or copy
everything with permissions, symlinks, etc (xxcopy?) if I can before it
fails totally.

Suggestions on best route to take when the replacement drive arrives.
Failing drive is SATA 250GB Hitachi partioned as
/dev/sda1 on /
/dev/sda2 on swap
/dev/sda6 on /home
/dev/sda5 on /usr
/dev/sda7 on /var

Replacement is 250Gb SATA Seagate

Regards
Gordon

Gordon Darling

ongelezen,
30 jan 2008, 09:49:4730-01-2008
aan

Further to above I've downloaded and burned g4l (Ghost for Linux) which
apparently does "click-n-burn".

However, I'll have to edit fstab as SuSE use disk-by-id

as in existing fstab

/dev/disk/by-id/scsi-SATA_Hitachi_HDT7250_VFA100R112282Kpart1 /
/dev/disk/by-id/scsi-SATA_Hitachi_HDT7250_VFA100R112282K-part6 /home
/dev/disk/by-id/scsi-SATA_Hitachi_HDT7250_VFA100R112282K-part5 /usr
/dev/disk/by-id/scsi-SATA_Hitachi_HDT7250_VFA100R112282K-part7 /var
/dev/disk/by-id/scsi-SATA_Hitachi_HDT7250_VFA100R112282K-part2 swap

Can I get away with editing it to/dev/sda1 / , etc or how do I obtain the
disk-id for the new disk.

Regards
Gordon

Don Raboud

ongelezen,
30 jan 2008, 10:45:0430-01-2008
aan
Gordon Darling wrote:

New disk is the same size as the failing one I presume?

If the failing drive is /dev/sda and the new dirve is /dev/sbd, try

dd if=/dev/sda of=/dev/sdb bs=8M conv=noerror

This will make an exact copy of the disk (including all partition
information) subject to reading errors. The conv=noerror tells dd to
continue if read errors are encountered.

You could also set up the partitions on the new disk before hand and use dd
once for each partition separately, like

dd if=/dev/sda1 of=/dev/sdb1 bs=8M conv=noerror
dd if=/dev/sda2 of=/dev/sdb2 bs=8M conv=noerror
...


Good luck.
--
Don

ray

ongelezen,
30 jan 2008, 11:24:4130-01-2008
aan

I believe you'll also need conv=sync - noerror keeps it going past
errors; sync replaces bad blocks with nulls - otherwise data gets
scrambled, at least that was my experience in backing up partitions to a
file.

ray

ongelezen,
30 jan 2008, 11:26:0130-01-2008
aan
On Wed, 30 Jan 2008 08:25:41 -0600, Gordon Darling wrote:

FWIW - Not too long ago I replace a 40gb disk in my laptop with a 120gb.
I backed everything to an external USB drive using partimage from a Live
CD - then swapped drives and restored them.

steve

ongelezen,
31 jan 2008, 16:43:4831-01-2008
aan
On Wed, 30 Jan 2008 22:49:47 +0800, Gordon Darling wrote
(in article <neydnYGMfIYWEz3a...@pipex.net>):

I presume you have extra cooling on your hard disk right?

Steve

0 nieuwe berichten