You will want to break the mdadm raid device and put each drive into your zfs array directly. That is one of the beautiful things about zfs
I knew I should have taken notes at the recent SILUG meeting when Steve
was showing us how this was done... :-( I want to convert a RAID5 array
from ext4 to zfs.
If you don't want to read a bunch of background info, jump directly to my
questions above my signature block.
----------------------------------------------------------------
I'm using a SuperMicro dual quad-Athlon motherboard with an Adaptec
zero-channel RAID card. I would like to use the on-board RAID firmware,
but it's old and limited to volumes under 2G. Instead I use software RAID
to merge nine 300GB SCSI3 drives on the Adaptec's two ports into a single
meta-drive (md127p1).
At the software level each drive looks like this:
# fdisk -l /dev/sdb
Disk /dev/sdb: 300.0 GB, 299965284352 bytes
255 heads, 63 sectors/track, 36468 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 512 bytes / 512 bytes
Disk identifier: 0x0005204c
Device Boot Start End Blocks Id System
/dev/sdb1 1 36468 292929178+ 83 Linux
At the hardware level they look like this:
md/raid:md127: device sdg operational as raid disk 5
md/raid:md127: device sdf operational as raid disk 4
md/raid:md127: device sdh operational as raid disk 6
md/raid:md127: device sdj operational as raid disk 8
md/raid:md127: device sdi operational as raid disk 7
md/raid:md127: device sdc operational as raid disk 1
md/raid:md127: device sdb operational as raid disk 0
md/raid:md127: device sde operational as raid disk 3
md/raid:md127: device sdd operational as raid disk 2
md/raid:md127: allocated 9574kB
md/raid:md127: raid level 5 active with 9 out of 9 devices,
algorithm 2
RAID conf printout:
--- level:5 rd:9 wd:9
disk 0, o:1, dev:sdb
disk 1, o:1, dev:sdc
disk 2, o:1, dev:sdd
disk 3, o:1, dev:sde
disk 4, o:1, dev:sdg
disk 5, o:1, dev:sdg
disk 6, o:1, dev:sdh
disk 7, o:1, dev:sdi
disk 8, o:1, dev:sdj
md127: detected capacity change from 0 to 2399712313344
md127: p1
I mention this because zfs' documentation seems to think the
number of SCSI channels is important. That said, I downloaded the
dkms-enabled ZFS package like Steve did and installed it on my
CentOS 6.4 server.
# yum localinstall --nogpgcheck
http://archive.zfsonlinux.org/epel/zfs-release-1-2.el6.noarch.rpm
# yum -y install zfs
Then I tried something simple (and intuitive):
# mkfs zfs /dev/md127p1
mk32fs 1.41.12 (17-May-2010)
mkfs.ext2: invalid blocks count - /dev/md127p1
Oops! I forgot that zfs isn't integrated into mkfs (yet). I seem to recall
a 'create' command. So I tried this:
# zfs create /dev/md127p1
Failed to load ZFS module stack.
Load the module manually by running 'insmod <location>/zfs.ko' as
root.
Failed to load ZFS module stack.
Load the module manually by running 'insmod <location>/zfs.ko' as
root.
That's promising, but there's no zfs.ko anywhere on my disk. Do I have to
make one. Something to do with dkms? Who knows?
--Doc Savage
Fairview Heights, IL
-
To unsubscribe, send email to majordomo@silug.org with
"unsubscribe silug-discuss" in the body.