[SATLUG] RAID5 Recovery - ddrescue 3rd round - CMP'd

Don Wright satlug at sbcglobal.net
Tue Aug 25 13:04:52 CDT 2009


On Tue, 25 Aug 2009 09:32:04 -0700 (PDT), David Labens
<adlabens at swbell.net> wrote:

>HOW do I confirm that all 4 disks have a proper SuperBlock?

mdadm --examine /dev/sda1 /dev/sdb1 /dev/sdc1 /dev/sde1

This is a command you've already been using. If there is no superblock,
the command will say so. (You could throw in /dev/sdd1 to see.) The
Array UUID will be the same on all members of your RAID array. The
Device UUID will be different. If any of those are the same, you've got
two copies of the same data installed, such as a drive and its ddrescue
clone.

You can verify the components (but not their order) with "fdisk -l",
which lists the partition tables of all installed drives, including
virtual drives from RAID and LVM. Don't worry about errors on those just
yet - they often omit the MSDOS partition tables fdisk is expecting.

The above mdadm command was based on your previous connections; if that
changed, use the ones you have now. The order of drive letters in a
software RAID array is not significant. The software figures it out from
the Array Slot number recorded in the superblock and shown in the
--examine output.

You should also check the Events count to be sure they are similar
before --force assembles the array. I'm concerned the obsolete data from
/dev/sdb1 (Events: 89) will pollute the mostly good data of the rest.

--Don (back but mostly working on computer show stuff.)

-- 
A computer is a hole in your desk you pour time into.


More information about the SATLUG mailing list