linux raid

Raid stands for Redundant Arrays o Independent Disks. It provides fault tolerance capability and / or performance enhancement depending on the level o RAID you are implementing.

step	Step	Raid- Linear	Raid-0 (danger)	Raid-1 (mirror)	Raid-4/5
make Partition	fdisk	fdisk -l /dev/hdb && list the disk information fdisk /dev/hdb && prepare a new partition it will go into fdisk prompt.   Option fdisk create partition n create a new partition   p Primary partition   2 2 second partion   Enter default clinder shown   + 200m 200M Bytes partition information t change system type of the parition   p print the partition information       Write to harddisk w
Raid information	vi /etc/raidtab		raid 0	raid 1	raid 5
	raiddev	/dev/md0	/dev/md0	/dev/mdo	/dev/md0
	raid-level	linear	0	1	5
	nr-raid-disks	2	2	2	3
	nr-spare-disks 後備	0	0	1	1
	persistent-superblock	0	0	0	1
	chunk-size	4	4	4	128
	device	/dev/hdb1	/dev/hdb1	/dev/hdb1	/dev/hdb1
	raid-disk	0	0	0	0
	device	/dev/hdd1	/dev/hdd1	/dev/hdd1	/dev/hdd1
	raid-disk	1	1	1	1
	device				/dev/hdc1
	raid-disk				2
	device			/dev/hdc1
	spare-disk			0
	raid-disk
Buid Raid (from information)	mkraid /dev/md0				mkraid --really-force /dev/md0
Check Raid	cat /proc/mdstat				cat /proc/mdstat&& the symbol "[UUU]" means that all three disks forming the RAID-5 are healthy
Format Raid	mke2fs /dev/md0				mke2fs -b 4096 -R strids=32 /dev/md0
Mount	mkdir /mnt/temp mount -t ext2 /dev/md0 /mnt/temp mount				mount -t ext2 /dev/md0 /mnt/temp
Check the Raid	df -m hdparm -t /dev/hdd1 /dev/md0
test it failure	raidsetfault /dev/md0 /dev/hdb1
/etc/fstab		/dev/md0 /mnt/temp ext2 defaults 0 0
Recover		replace a new one	replace
		modify /etc/raidtab	modify /etc/raidtab	raidhotremove /dev/md0 /dev/hdd1	raidhotremove /dev/md0 /dev/hdd1
		invoice mkraid /dev/md0	invoke mktraid /dev/md0	raidhotadd /dev/md0 /dev/hdd1	raidhotadd /dev/md0 /dev/hdd1
				cat /proc/dstat && view condition	cat /proc/mdstat