Server Fault is a question and answer site for system and network administrators. It's 100% free, no registration required.

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I have two 500GB hard drives in my server. Currently, there is no RAID configured, but CentOS6 is installed (as a bare install for now). Is it possible to set up a software RAID-1 so that both drives are mirrored without the need to reinstall the OS? I have access to KVM for emergencies, but I'd prefer to be able to do it all through SSH.

share|improve this question
If this is a professional situation versus a learning opportunity or curiosity, I'd just reinstall and configure your software RAID correctly. – ewwhite Nov 9 '13 at 17:19

You can do this (which I never tried myself, so test your KVM access first!):

First device to mirror either entire disks (/dev/sdX) or just partitions. Even if you only mirror a single full disk partition. In the examples below I assumed whole disk mirror.

mdadm --create /dev/md0 -n 2 -l 1 missing /dev/sdb
This creates a RAID-1 mirror with one disk missing.

Copy all the data from your first disk to the array.
Rsync might be useful for this. Exclude /proc and /dev in your copy.
Partitions might need to be created. There is not enough information in your post to indicate if this is the case or not.

Set up a proper boot on the new MD device. Keep an option in grub2 (or what your boot manager is) to fall back to the old disk in case it does not work.

Reboot to the new disk. Ignore the degraded array state warning.

Add the old disk to the array mdadm --manage /dev/md0 -a /dev/sda and let it synchronise.

[Edit] Please do report back on this if you needed anything extra. E.g. a modprobe raid1 is CentOS did not come with the mirror module loaded by default. As written at the beginning of the answer: UNTESTED).

Actually tested stuff:

Step 1: Create the array

I created a VM with two 5 GB disks in Vmware workstation. I downloaded the CentOS 6.4 ISO and installed it on the first disk using a single partition. Maybe not the best way to partition a disk, but this is for a test only.

As you can see there is only one disk in use after booting:

[root@centOS-RAID-test etc]# mount
/dev/sda1 on / type ext4 (rw)
proc on /proc type proc (rw)
sysfs on /sys type sysfs (rw)
devpts on /dev/pts type devpts (rw,gid=5,mode=620)
tmpfs on /dev/shm type tmpfs (rw,rootcontext="system_u:object_r:tmpfs_t:s0")
none on /proc/sys/fs/binfmt_misc type binfmt_misc (rw)

[root@centOS-RAID-test etc]# cat fstab
# /etc/fstab
# Created by anaconda on Sun Nov 10 01:19:26 2013
# Accessible filesystems, by reference, are maintained under '/dev/disk'
# See man pages fstab(5), findfs(8), mount(8) and/or blkid(8) for more info
# UUID=ccb58393-d92e-473f-ae8d-7b2d7231dae8 /   ext4    defaults        1 1
/dev/sda1               /                       ext4    defaults        1 1
tmpfs                   /dev/shm                tmpfs   defaults        0 0
devpts                  /dev/pts                devpts  gid=5,mode=620  0 0
sysfs                   /sys                    sysfs   defaults        0 0
proc                    /proc                   proc    defaults        0 0

I made one change from the default and that is from the UUID based mount-point to /dev/sda1. I did this since I think it is easier to identify the disks this way. I rebooted after the change to make sure I did not somehow broke the system.

Next lets use mdadm.

curl > file
  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed
100  333k  100  333k    0     0   339k      0 --:--:-- --:--:-- --:--:-- 1044k

[root@centOS-RAID-test ~]# rpm -ivh file
warning: file: Header V3 RSA/SHA1 Signature, key ID c105b9de: NOKEY
Preparing...                ########################################### [100%]
   1:mdadm                  ########################################### [100%]

[root@centOS-RAID-test ~]# mdadm
Usage: mdadm --help
  for help

Ok, the mdadm command seems to be present and in my path. No need to rehash.

[root@centOS-RAID-test ~]# mdadm --create /dev/md0 -n 2 -l 1 missing /dev/sdb
mdadm: /dev/sdb appears to be part of a raid array:
    level=raid0 devices=0 ctime=Thu Jan  1 01:00:00 1970
mdadm: partition table exists on /dev/sdb but will be lost or
       meaningless after creating array
mdadm: Note: this array has metadata at the start and
    may not be suitable as a boot device.  If you plan to
    store '/boot' on this device please ensure that
    your boot-loader understands md/v1.x metadata, or use
Continue creating array? y
mdadm: Defaulting to version 1.2 metadata
mdadm: array /dev/md0 started.

[root@centOS-RAID-test ~]# cat /proc/mdstat
Personalities : [raid1]
md0 : active raid1 sdb[1]
      5238720 blocks super 1.2 [2/1] [_U]

I am not sure why it thinks that this unused disk was part of a previous array, but the new md device successfully gets created without any 'Device or Resource busy' error.

Notice the name: MD0. When I rebooted this changed to md127.

To keep this consistent create /etc/mdadm.conf. I used ARRAY /dev/md0 level=raid1 num-devices=2 metadata=1.2 devices=/dev/sdb as content. Some googling shows that this can be automated with mdadm --verbose --detail --scan > /etc/mdadm.conf.

Stopped updating since it is almost 3AM. Will continue tomorrow

share|improve this answer
Given this a try, but I'm getting 'Device or Resource busy' when running the mdadm create command. This makes sense since the /dev/sda has mounted partitions (/boot and /). Is there a workaround for this? – Alex Blundell Nov 9 '13 at 0:32
Use the missing part for the disk which is already active and only used the empty unmounted second disk to initially create the array. In your case that might be /dev/sdb. – Hennes Nov 9 '13 at 0:34
Hmm.. I'm getting the same result for both devices. /dev/sdb is definitely the unmounted extra disk, so I've put that instead of sda, but I'm still getting device busy. It's not mounted either (I've checked /proc/mounts) – Alex Blundell Nov 9 '13 at 0:39
I failed to reproduce the error. See the partially updated post. – Hennes Nov 10 '13 at 1:49

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.