mdadm -c magic?

mdadm -c magic?

[Robert L. Harris]

[-- Attachment #1: Type: text/plain, Size: 1837 bytes --]


Situation, I need to get a RAID5 array online ASAP.  It's to be 4 160Gig IDE
drives on a 3ware raid controller.

One hitch, I just lit up the box and started to mess with the disks to
make the array and it seems one disk is having "issues"...  I'm sending
it off for RMA which can take 1-2 weeks easily.

Is it possible to have mdadm make the array and just leave the one spot
open anyway?  I tried a few options:

root@legato-disk5.acs:~# mdadm -C -c 128  -l 5 -p left-asymmetric -n 4 -x 0=
 /dev/md3 /dev/sdm1 /dev/sdn1 /dev/sdo1 /dev/sdp1
mdadm: /dev/sdm1 appears to contain an ext2fs file system
    size=3D480238464K  mtime=3DTue Dec  3 11:11:22 2002
mdadm: /dev/sdm1 appear to be part of a raid array:
    level=3D5 disks=3D4 ctime=3DTue Dec  3 10:48:56 2002
mdadm: /dev/sdo1 appear to be part of a raid array:
    level=3D5 disks=3D4 ctime=3DTue Dec  3 10:48:56 2002
mdadm: Cannot open /dev/sdp1: No such file or directory
mdadm: create aborted
root@legato-disk5.acs:~# mdadm -C -c 128  -l 5 -p left-asymmetric -n 4 -x 0=
 /dev/md3 /dev/sdm1 /dev/sdn1 /dev/sdo1
mdadm: You haven't given enough devices (real or missing) to create this ar=
ray


I figured trying to create it with 3 drives and then adding a 4th would
just make that 4th a spare.  Yes I know it's not very redundant with 3
of 4 drives but atleast I can get it online and start testing with light
data load.  If it fails another disk I won't be any farther behind than
if I just waited...

Robert


:wq!
---------------------------------------------------------------------------
Robert L. Harris                     | PGP Key ID: E344DA3B
                                         @ x-hkp://pgp.mit.edu=20
DISCLAIMER:
      These are MY OPINIONS ALONE.  I speak for no-one else.
FYI:
 perl -e 'print $i=3Dpack(c5,(41*2),sqrt(7056),(unpack(c,H)-2),oct(115),10)=
;'


[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: mdadm -c magic?

[Derek Vadala]

On Thu, 20 Feb 2003, Robert L. Harris wrote:

> Is it possible to have mdadm make the array and just leave the one spot
> open anyway?  I tried a few options:
>
> root@legato-disk5.acs:~# mdadm -C -c 128  -l 5 -p left-asymmetric -n 4 -x 0=
>  /dev/md3 /dev/sdm1 /dev/sdn1 /dev/sdo1 /dev/sdp1
> [snip]
> mdadm: create aborted
> root@legato-disk5.acs:~# mdadm -C -c 128  -l 5 -p left-asymmetric -n 4 -x 0=
>  /dev/md3 /dev/sdm1 /dev/sdn1 /dev/sdo1
> mdadm: You haven't given enough devices (real or missing) to create this ar=
> ray

Try using the keyword 'missing' in place of a real disk or partition:

# mdadm -C /dev/md3 -c 128 -l 5 -n 4 /dev/sdm1 /dev/sdn1 /dev/sdo1 missing

That should give you a 3-disk RAID-5 in degraded mode. I think that's what
you want. Add the fourth disk when it shows up and the array will
resync.

Re: mdadm -c magic?

[Robert L. Harris]

[-- Attachment #1: Type: text/plain, Size: 1347 bytes --]



Bingo.  Thank you very much.


Thus spake Derek Vadala (derek@vadala.com):

> On Thu, 20 Feb 2003, Robert L. Harris wrote:
> 
> > Is it possible to have mdadm make the array and just leave the one spot
> > open anyway?  I tried a few options:
> >
> > root@legato-disk5.acs:~# mdadm -C -c 128  -l 5 -p left-asymmetric -n 4 -x 0=
> >  /dev/md3 /dev/sdm1 /dev/sdn1 /dev/sdo1 /dev/sdp1
> > [snip]
> > mdadm: create aborted
> > root@legato-disk5.acs:~# mdadm -C -c 128  -l 5 -p left-asymmetric -n 4 -x 0=
> >  /dev/md3 /dev/sdm1 /dev/sdn1 /dev/sdo1
> > mdadm: You haven't given enough devices (real or missing) to create this ar=
> > ray
> 
> Try using the keyword 'missing' in place of a real disk or partition:
> 
> # mdadm -C /dev/md3 -c 128 -l 5 -n 4 /dev/sdm1 /dev/sdn1 /dev/sdo1 missing
> 
> That should give you a 3-disk RAID-5 in degraded mode. I think that's what
> you want. Add the fourth disk when it shows up and the array will
> resync.

:wq!
---------------------------------------------------------------------------
Robert L. Harris                     | PGP Key ID: E344DA3B
                                         @ x-hkp://pgp.mit.edu 
DISCLAIMER:
      These are MY OPINIONS ALONE.  I speak for no-one else.
FYI:
 perl -e 'print $i=pack(c5,(41*2),sqrt(7056),(unpack(c,H)-2),oct(115),10);'


[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]