On Wed, 28 Jun 2000, Chris Quinn wrote:

>Hi,
>
>Can anyone suggest a solution to this seemingly simple problem, apart from 
>labotomising myself :o) ?
>I had hdaX and hdbX devices mirrored under raidtools 0.9.
>I moved the hdb disk physically to hdc and it was no longer recognised as hda's 
>counterpart. My raid array thus runs in degraded mode. It does not allow me to go 
>back to hdb either.
>Instead I tried binary copying hda to hdc and looking for the raid command to force 
>hdcX's into the arrays as second devices.
>I remember being able to do this under a previous version but 
>cannot seem to do it under 0.9. Indeed the manpage on mkraid
>states that initialising a device destroys the data.
>
>I tried changing /etc/raidtab for one of the arrays:
>
>raiddev             /dev/md6
>raid-level                  1
>nr-raid-disks               2
>chunk-size                  64k
>persistent-superblock       1
>    device          /dev/null
>    raid-disk     0
>    device          /dev/hdc11
>    raid-disk     1                                                             
>
>...then ran mkraid --really-force /dev/md6.
>Mkraid complained about /dev/null.
>Next I went down to one device with 'raid-disk 1' and it
>complained, so I  made it raid-disk 0. This of course worked but I didn't want the 
>superblock recording itself as the principal disk.
>
>There seemed to be rebuild activity at this point despite the fact that the array was 
>composed of only one disk.
>What I do not want to do now is bring the machine back up with
>hda11 and hdc11 both claiming to be the main device and have hda11 rebuilt with 
>non-existent data from hdc11 !!
>
>Does anyone know a way round this?
>I imagined adding a fresh/spare disk to a mirroring group (even if running in 
>degraded mode) would be a very basic task!
>Surely I don't have to recreate a new empty 2-device array and fill it with the data 
>from hda11?
>
>Any words of wisdom welcome...
>Thanks
>Chris Q.
>


You might want to run ``raidhotadd /dev/md? /dev/hdc?''
I had the same problem, when I ``accidently'' added an other scsi disk (in
between) to my system..
degraded mode: cat /proc/mdstat gave somthing like ``... [U_]''. I solved
it with hotadd. basicly it just looks like It has droped a disk, so add a
new one (no difference if it's the same 8).

Yours
--
  Morten Bøgeskov (email: [EMAIL PROTECTED])

Drunk as I may be, I've still got 2 cents.

Reply via email to