Resources
http://www.tldp.org/HOWTO/Software-RAID-HOWTO.html#toc5
http://www.linuxdevcenter.com/pub/a/linux/2002/12/05/RAID.html
http://linux-raid.osdl.org/index.php/RAID_setup
Related: LVM
http://www.tldp.org/HOWTO/LVM-HOWTO/
http://contribs.org/modules/phpwiki/index.php/lvm
SME 7
SME 7 installation on a single disk defaults to configuring a degraded RAID1 (mirrored) array, on two disks, RAID1, on three disks, RAID5.
A second disk added to a single disk SME server will not automatically create the mirror.
SME 7 has a new option in admin menu to "Manage disk redundancy". Here you can add a second hard disk to complete the mirror.
su admin
SME 7 uses the raidmonitor package to alert the admin user of raid problems.
Useful RAID Commands
View partitions:
cat /proc/partitions
View current status of all arrays:
cat /proc/mdstat
View array configuration file:
cat /etc/mdadm.conf
View array details:
mdadm --detail --scan mdadm --detail /dev/mdx
Remove the problem partition from array:
mdadm /dev/mdx -r /dev/sdxy
Add the partition to the array:
mdadm /dev/mdx -a /dev/sdxy
Reset the raidmonitor 'last known good configuration' after raid repairs are complete:
raidmonitor -iv
You may even need to run this command again after a reboot, if ALERT! messages resume.
Alert messages are sent to " This e-mail address is being protected from spambots. You need JavaScript enabled to view it ", which is an alias (Pseudonym) that can be redirected in the Server-Manager.
Reconstruction - Basic Steps - Same Disk or Partition
- Check the array configuration
- cat /proc/mdstat
- Verify the details of the failed volume
- mdadm --detail /dev/mdx
- Remove the failed partition
- mdadm --remove /dev/mdx /dev/sdxy
- Add the same partition as a hot spare
- mdadm --add /dev/mdx /dev/sdxy
- Watch the automatic reconstruction begin
- cat /proc/mdstat
Reconstruction - Basic Steps - Disk Replaced
- Power down the system
- Replace the failed disk
- Power up the system once again.
- Use "mdadm /dev/mdx -a /dev/sdxx" to re-insert the disk partition in the array
- Watch the automatic reconstruction running
Disk Repair
This command will scan your disk and, optionally, mark off the bad blocks on it.
badblocks -vf /dev/hdX # test only badblocks -wvf /dev/hdX # write test and remap bad blocks
Only run this on unmounted disks. It takes a LONG time to run.
Speed Up RAID1 Resync
Just echo large numbers into both min and max resync speed proc entries: /proc/sys/dev/raid/spped_limit_max and speed_limit_min
As root do:
echo 500000 > /proc/sys/dev/raid/speed_limit_max echo 500000 > /proc/sys/dev/raid/speed_limit_min
The above will set both speed limits to 500MiB/s which ought to be more than your devices can do...
Remove Arrays
As root do:
- Determine which array (md0, md1, etc..) using fdisk -l
- Make sure the array to be deleted is unmounted using umount /dev/mdX
- Query your arrays to find out what disks are contained using mdadm --detail /dev/mdX
- Shut down the array using mdadm --stop /dev/mdX
- Zero the superblock for each partition in the array using mdadm --zero-superblock /dev/sdXY
| < Prev | Next > |
|---|
