how do I change md1 to md2?

how do I change md1 to md2?

[David Nedved]

Hi All,

Thanks for everyone who helped me get my md1 data volume brought up
correctly.  Now I have another issue.  I've rebuilt my host OS using
RedHat 7.3, and I loaded it on RAID1 also.  I built my data volume as
md1 so that I could load the OS on md0, but I forgot about swap.

Now I have md0 as /, md1 as swap, and another md1 data volume that
won't activate since there is already an md1 in use as swap.

Can anyone help me get my data volume brought up as md2?  I have it in
the raidtab as md2, but during boot it reads the superblock and knows it's
md1, but won't bring it up since there's already an md1.  I've gone back
to the old OS disk, so I don't have logs handy, but if they're critical it
won't be a problem to post them.  I looked byt coulnd't find any pointer
in the FAQ or Howto's...

I know I could probably turn swap off, so that the data volume activates
correctly, but I'm considering putting software RAID1 onto production
servers, and now that I've found this issue, I really feel I need to
understand how to manage it rather than work around it.

Thanks in advance,

David

Re: how do I change md1 to md2?

[tc lewis]

i believe you can use neil brown's mdadm utility to update superblock
data:
http://www.cse.unsw.edu.au/~neilb/source/mdadm/

according to a recent mail in the archive, version 1.0.0 is recommended
and 1.0.1 is experimental.  i've never used the utility myself.

i think there was also an old trick to adjust your raidtab, but change one
of your "raid-disk"s to a "failed-disk", in the config for the proper md
device entry, and run mkraid --force /dev/md?, and that would write a new
superblock yet not erase all the data (as long as it thought there was a
failed disk)...but you should probably get confirmation / more details on
that first if you're going to try that method, to prevent losing data.

you may also be able to add arguments to your kernel loader, re: with lilo
or whatever.  Documentation/md.txt in the linux source has info on that.
that may work to override a device # for that boot, but it won't fix the
superblock permanently.

i'm sure there are other methods too.  i'm not sure what the most
recommended option is these days -- mdadm, perhaps.

search a linux-raid archive for more info on mdadm, or its previous name,
mdctl.

-tcl.


On Mon, 10 Jun 2002, David Nedved wrote:

> Hi All,
>
> Thanks for everyone who helped me get my md1 data volume brought up
> correctly.  Now I have another issue.  I've rebuilt my host OS using
> RedHat 7.3, and I loaded it on RAID1 also.  I built my data volume as
> md1 so that I could load the OS on md0, but I forgot about swap.
>
> Now I have md0 as /, md1 as swap, and another md1 data volume that
> won't activate since there is already an md1 in use as swap.
>
> Can anyone help me get my data volume brought up as md2?  I have it in
> the raidtab as md2, but during boot it reads the superblock and knows it's
> md1, but won't bring it up since there's already an md1.  I've gone back
> to the old OS disk, so I don't have logs handy, but if they're critical it
> won't be a problem to post them.  I looked byt coulnd't find any pointer
> in the FAQ or Howto's...
>
> I know I could probably turn swap off, so that the data volume activates
> correctly, but I'm considering putting software RAID1 onto production
> servers, and now that I've found this issue, I really feel I need to
> understand how to manage it rather than work around it.
>
> Thanks in advance,
>
> David
> -
> 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
>