alexalmazan.com

Establishing software RAID is done through the use of ‘mdadm’ via the CLI

mdadm -C ‘RAID dev’ -l ‘level’ -n ’#of disks/partitions’

Here is an example designating device /dev/md0 set to RAID 1 amongst the partitions /dev/sda5 and /dev/sda6

mdadm --create /dev/md0 -l 1 -n 2 /dev/sda5 /dev/sda6

once done you can check the status in real time with

cat /proc/mdstat

[root@server1]# cat /proc/mdstat%

personalities : {raid1} md0 : active raid1

Once the RAID is active, it requires a file system and label. Create a file system and label on the new container with the following syntax

[root@server1]#mke2fs -j /dev/md0 -L raid1

Next, get the device mounted and available upon the next server reboot with

[root@server1]_tail -i /etc/mtab >> /etc/fstab 

This will add to the /etc/fstab file or manually edit /etc/fstab with something similar to

/dev/md0    /mnt    ext3    rw    0 0

The following command sequence using ‘mdadm’ fails drive in the array-

mdadm /dev/md0 -f /dev/sda5

removes failed drive

[root@server1]#mdadm /dev/md0 -r /dev/sda5

re adds failed drive

mdadm /dev/md0 -a /dev/sda5 

mdadm—detail

FDISK RAID example

fdisk /dev/sda n

create extended

select all sizes n

once the partitions are written, ‘partprobe’

(ex. device creation in /dev/) mknod sda6 b 8 6

mdadm _X

DONT forget to issue ‘partprobe’ after fdisk

mdadm -C /dev/md1 -l 5 -n 3 -x 1 /dev/sda{7.8.9.10}

n is the number of components -x spare components then fail a drive with

mdadm /dev/md1 -f /dev/sda7 

watch -i /cat/proc/mdstat to see the failure