ADD_NEW_DISK failed

ADD_NEW_DISK failed

[Victor Geller]

Hi,

I have the following problem:

[root@wti7 vic]# mdadm -Cv /dev/md0 -l1 -n2 /dev/hda1 /dev/hdc1
mdadm: chunk size defaults to 64K
mdadm: /dev/hda1 appears to contain an ext2fs file system
     size=5630748K  mtime=Sun Feb 23 08:52:48 2003
mdadm: /dev/hdc1 appears to contain an ext2fs file system
     size=5630748K  mtime=Sun Feb 23 08:52:54 2003
mdadm: size set to 5630656K
Continue creating array? y
mdadm: ADD_NEW_DISK for /dev/hda1 failed: Invalid argument

[root@wti7 vic]# cat /proc/mdstat
Personalities :
read_ahead not set
md0 : inactive
unused devices: <none>

Partition and mount info is below:

[root@wti7 vic]# fdisk -ul /dev/hda

Disk /dev/hda: 120.0 GB, 120000000000 bytes
255 heads, 63 sectors/track, 14589 cylinders, total 234375000 sectors
Units = sectors of 1 * 512 = 512 bytes

    Device Boot    Start       End    Blocks   Id  System
/dev/hda1   *        63  11261564   5630751   83  Linux
/dev/hda2      11261565 234372284 111555360    5  Extended
/dev/hda5      11261628  11759579    248976   82  Linux swap
/dev/hda6      11759643 234372284 111306321   83  Linux

[root@wti7 vic]# fdisk -ul /dev/hdc

Disk /dev/hdc: 120.0 GB, 120000000000 bytes
255 heads, 63 sectors/track, 14589 cylinders, total 234375000 sectors
Units = sectors of 1 * 512 = 512 bytes

    Device Boot    Start       End    Blocks   Id  System
/dev/hdc1   *        63  11261564   5630751   83  Linux
/dev/hdc2      11261565 234372284 111555360    5  Extended
/dev/hdc5      11261628  11759579    248976   82  Linux swap
/dev/hdc6      11759643 234372284 111306321   83  Linux

[root@wti7 vic]# mount
/dev/hda1 on / type ext3 (rw)
none on /proc type proc (rw)
none on /proc/bus/usb type usbdevfs (rw)
none on /dev type devfs (rw)
none on /dev/pts type devpts (rw,mode=0620)
/dev/hda6 on /home type ext3 (rw)
none on /mnt/cdrom type supermount 
(ro,dev=/dev/scd0,fs=auto,--,iocharset=iso885
9-1,codepage=850,umask=0)
none on /mnt/floppy type supermount 
(rw,sync,dev=/dev/fd0,fs=auto,--,iocharset=i
so8859-1,codepage=850,umask=0)
/dev/hdc1 on /mnt/raid1_1 type ext3 (rw)
/dev/hdc6 on /mnt/raid_1/home type ext3 (rw)

I would appreciate your help

Victor Geller , vgeller@spindrift.on.ca

Re: ADD_NEW_DISK failed

[Norman Schmidt]

Hi Victor!

Victor Geller schrieb:
> Hi,
> 
> I have the following problem:
> 
> [root@wti7 vic]# mdadm -Cv /dev/md0 -l1 -n2 /dev/hda1 /dev/hdc1
> mdadm: chunk size defaults to 64K
> mdadm: /dev/hda1 appears to contain an ext2fs file system
>     size=5630748K  mtime=Sun Feb 23 08:52:48 2003
> mdadm: /dev/hdc1 appears to contain an ext2fs file system
>     size=5630748K  mtime=Sun Feb 23 08:52:54 2003
> mdadm: size set to 5630656K
> Continue creating array? y
> mdadm: ADD_NEW_DISK for /dev/hda1 failed: Invalid argument
> 
> [root@wti7 vic]# mount
> /dev/hda1 on / type ext3 (rw)
> /dev/hdc1 on /mnt/raid1_1 type ext3 (rw)
> 
> I would appreciate your help
> 
> Victor Geller , vgeller@spindrift.on.ca

That is easy:

You MUST unmount /dev/hda1 and /dev/hdc1 before using them as part of a 
md device! Later, you will NEVER mount one of both but ONLY /dev/md0 on /.
But this will not work with your current configuration, since you 
actually cannot unmount the root fs while it is running.
Ask Neil or whoever how to use the failed disk way for circumventing the 
need to copy your rootfs twice to another place (forth and back) or 
newly install the system (I am preparing a modified Debian Woody 
bootable CD just for that).

By the way: This failed-disk approach is not so well documented for 
mdadm, or did I overlook something? Any documentation anywhere?

Hth, Norman.

-- 
--

Norman Schmidt          Institut für Physikal. u. Theoret. Chemie
cand. chem.             Friedrich-Alexander-Universitaet
schmidt@naa.net         Erlangen-Nuernberg

-
To unsubscribe from this list: send the line "unsubscribe linux-raid" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: ADD_NEW_DISK failed

[Neil Brown]

On Sunday February 23, norman.schmidt@ratnet.stw.uni-erlangen.de wrote:
> 
> By the way: This failed-disk approach is not so well documented for 
> mdadm, or did I overlook something? Any documentation anywhere?
> 

Fourth paragraph of the "CREATE MODE" section:

       To create a "degraded" array in which some devices are missing,  simply
       give the word missing in place of a device name.  This will cause mdadm
       to leave the corresponding slot in the array empty.   For  a  RAID4  or
       RAID5  array  at most one slot can be missing.  For a RAID1 array, only
       one real device needs to be given.  All of the others can be missing.


If that is not clear, please help me re-phrase it.


NeilBrown

Re: ADD_NEW_DISK failed

[Adam Luter]

Just add "" around missing:

... give the word "missing" in place ...

-Adam Luter

On Mon, Feb 24, 2003 at 07:35:36AM +1100, Neil Brown wrote:
> Fourth paragraph of the "CREATE MODE" section:
> 
>        To create a "degraded" array in which some devices are missing,  simply
>        give the word missing in place of a device name.  This will cause mdadm
>        to leave the corresponding slot in the array empty.   For  a  RAID4  or
>        RAID5  array  at most one slot can be missing.  For a RAID1 array, only
>        one real device needs to be given.  All of the others can be missing.

Re: ADD_NEW_DISK failed

[Victor Geller]

Norman, Neil and Adam,

Thank you for your help.

I am a bit confused on the sequence of steps I have to do.
I can
1. umount /dev/hdc*
2. mdadm -Cv /dev/md0 -n2 -l1 /dev/hdc1 "missing"
3. mdadm -Cv /dev/md1 -n2 -l1 /dev/hdc6 "missing"

What would be the next steps?


At 07:35 AM 2/24/2003 +1100, Neil Brown wrote:
>On Sunday February 23, norman.schmidt@ratnet.stw.uni-erlangen.de wrote:
> >
> > By the way: This failed-disk approach is not so well documented for
> > mdadm, or did I overlook something? Any documentation anywhere?
> >
>
>Fourth paragraph of the "CREATE MODE" section:
>
>        To create a "degraded" array in which some devices are 
> missing,  simply
>        give the word missing in place of a device name.  This will cause 
> mdadm
>        to leave the corresponding slot in the array 
> empty.   For  a  RAID4  or
>        RAID5  array  at most one slot can be missing.  For a RAID1 array, 
> only
>        one real device needs to be given.  All of the others can be missing.
>
>
>If that is not clear, please help me re-phrase it.
>
>
>NeilBrown

Re: ADD_NEW_DISK failed

[Neil Brown]

On Sunday February 23, luterac@bellsouth.net wrote:
> Just add "" around missing:
> 
> ... give the word "missing" in place ...

Well, it's bold in the original - cut-and-paste from an xterm lost
that.
Maybe quotes would make it clearer...

thanks,
NeilBrown

> 
> -Adam Luter
> 
> On Mon, Feb 24, 2003 at 07:35:36AM +1100, Neil Brown wrote:
> > Fourth paragraph of the "CREATE MODE" section:
> > 
> >        To create a "degraded" array in which some devices are missing,  simply
> >        give the word missing in place of a device name.  This will cause mdadm
> >        to leave the corresponding slot in the array empty.   For  a  RAID4  or
> >        RAID5  array  at most one slot can be missing.  For a RAID1 array, only
> >        one real device needs to be given.  All of the others can be missing.
> -
> To unsubscribe from this list: send the line "unsubscribe linux-raid" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: ADD_NEW_DISK failed

[Derek Vadala]

On Mon, 24 Feb 2003, Neil Brown wrote:

> Fourth paragraph of the "CREATE MODE" section:
>
>        To create a "degraded" array in which some devices are missing,  simply
>        give the word missing in place of a device name.  This will cause mdadm
>        to leave the corresponding slot in the array empty.   For  a  RAID4  or
>        RAID5  array  at most one slot can be missing.  For a RAID1 array, only
>        one real device needs to be given.  All of the others can be missing.

Perhaps an example immediately following this one or in the examples
section at the end of the manual page.

I hate to advocate something that rewards people for not RTFMing, but
since this is probably the single most asked question about mdadm, perhaps
some extra interactive logic would be helpful for this case.

The obvious,

	if --raid-devices > # of devices listed on the command line
		then ask if the user is trying to create a degraded array

seems like it might create more havoc for special cases though. I'm not
sure it's really a good alternative. Maybe it's just a good idea to refer
people to the section in the man page. Example:

	n.b.: For help with degraded arrays, see the corresponding entry
	in the manual page.


--
Derek Vadala, derek@cynicism.com, http://www.cynicism.com/~derek