We’ll be setting up a Linux software RAID10 and re-configuring MySQL to use it for data files and databases.
Installation
We’re using Ubuntu 13.10 (Saucy Salamander):
# uname -rv 3.11.0-15-generic #23-Ubuntu SMP Mon Dec 9 18:17:04 UTC 2013
Install mdadm and mysql:
# apt-get update && apt-get install mdadm mysql-server
Create Software RAID10
We have 4x2GB disks pre-configured for Linux raid:
# fdisk -l | grep /dev/xv Disk /dev/xvda1 doesn't contain a valid partition table Disk /dev/xvda1: 8589 MB, 8589934592 bytes Disk /dev/xvdf: 2147 MB, 2147483648 bytes /dev/xvdf1 2048 4194303 2096128 fd Linux raid autodetect Disk /dev/xvdg: 2147 MB, 2147483648 bytes /dev/xvdg1 2048 4194303 2096128 fd Linux raid autodetect Disk /dev/xvdh: 2147 MB, 2147483648 bytes /dev/xvdh1 2048 4194303 2096128 fd Linux raid autodetect Disk /dev/xvdi: 2147 MB, 2147483648 bytes /dev/xvdi1 2048 4194303 2096128 fd Linux raid autodetect
Create the RAID10 array:
# mdadm -C /dev/md0 -l 10 -n 4 /dev/xvdf1 /dev/xvdg1 /dev/xvdh1 /dev/xvdi1 -v mdadm: layout defaults to n2 mdadm: layout defaults to n2 mdadm: chunk size defaults to 512K mdadm: size set to 2094592K mdadm: Defaulting to version 1.2 metadata mdadm: array /dev/md0 started.
Scan and activate the array automatically on startup:
# mdadm --detail --scan --verbose > /etc/mdadm/mdadm.conf
Wait for the array to finish synchronising disks:
# watch cat /proc/mdstat Personalities : [linear] [multipath] [raid0] [raid1] [raid6] [raid5] [raid4] [raid10] md0 : active raid10 xvdi1[3] xvdh1[2] xvdg1[1] xvdf1[0] 4189184 blocks super 1.2 512K chunks 2 near-copies [4/4] [UUUU] [=========>...........] resync = 49.2% (2064912/4189184) finish=0.9min speed=38738K/sec unused devices: <none>
Synchronisation is done:
# cat /proc/mdstat Personalities : [linear] [multipath] [raid0] [raid1] [raid6] [raid5] [raid4] [raid10] md0 : active raid10 xvdi1[3] xvdh1[2] xvdg1[1] xvdf1[0] 4189184 blocks super 1.2 512K chunks 2 near-copies [4/4] [UUUU] unused devices: <none>
Note that it’s good practice to regularly run data scrubbing to check for and fix errors. To initiate a data scrub manually:
# echo check > /sys/block/md0/md/sync_action
To stop a currently running data scrub safely:
# echo idle > /sys/block/md0/md/sync_action
Create an ext4 filesystem:
# mkfs.ext4 /dev/md0
Create a mountpoint:
# mkdir /sql_data
Mount the raid drive:
# mount /dev/md0 /sql_data
Create a lost+found directory:
# cd /sql_data
# mklost+found
Don’t forget to add a record to fstab:
# echo "/dev/md0 /sql_data ext4 defaults 0 0" >> /etc/fstab
Optionally, configure email alerts for daily checks to ensure that any degraded MD devices don’t go unnoticed:
# echo "MAILADDR [email protected]" >> /etc/mdadm/mdadm.conf # echo "MAILFROM [email protected]" >> /etc/mdadm/mdadm.conf
Configure MySQL
By default, MySQL stores its data files and databases under /var/lib/mysql on Ubuntu. We want to change it to our newly created raid drive. Stop the daemon first:
# service mysql stop
Open the configuration file and change datadir appropriatelly:
# vim /etc/mysql/my.cnf [...] #datadir = /var/lib/mysql datadir = /sql_data [...]
Create an alias for the new path:
# echo "alias /var/lib/mysql/ -> /sql_data/," >> /etc/apparmor.d/tunables/alias
Copy all content form the default directory to the new location and change ownership of the folder:
# cp -Rp /var/lib/mysql/* /sql_data # chown mysql:mysql /sql_data # chmod 0700 /sql_data
Reload apparmor daemon and start mysql server:
# service apparmor reload # service mysql start
Check mysql logs in case you run into problems:
# tail -f /var/log/mysql/error.log