Quoting Dexter Filmore <Dexter.Filmore@xxxxxx>:
While I'm at migrating data to a new array I thought I get rid of LVM and
partition the array directly - spares me one extra layer that I don't need.
Now, I can cfdisk md1 just fine and sfdisk -d shows me the partitions
md1p[1-4], but as soon as I want to mkfs, I get:
-------------
# LC_ALL=en_EN mkfs.ext3 -O dir_index -E stride=16 -m 0 /dev/md1p1
mke2fs 1.40-WIP (14-Nov-2006)
Could not stat /dev/md1p1 --- No such file or directory
The device apparently does not exist; did you specify it correctly?
-------------
And lo and behold: in /dev no trace of md1p1 or md1/p1 or whatever. Only md1.
When assembling I get:
# mdadm -A -s
mdadm: that --auto option not compatable with device named /dev/md1
Here's my mdadm.conf:
DEVICE partitions
ARRAY /dev/md1 level=raid5 num-devices=5 auto=mdp
UUID=f2352335:f8654106:19a59613:120d8f39
This is debian 4.0r3 with kernel 2.6.22 from etch-backports and mdadm 2.5.6.
What do I have here? Version conflict?
It should work, but I *sorta* remember there might be a bug in the
raid device name for that version of mdadm.
The quick fix would be name your raid device /dev/md_d0 .
After its created, try partitioning it and rebooting.
You should hopefully then be able to mkfs your /dev/md_d0p1 and
/dev/md_d0p2 and so on.
Cheers,
Mike
--
To unsubscribe from this list: send the line "unsubscribe linux-raid" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
