raid
From WebHostingNeeds.com
- mdadm
- check raid health
- windows raid
- Hardware RAID
- http://www.thecloudcalculator.com/calculators/disk-raid-and-iops.html
- RAID 1 and RAID 10 are your only really decent options. Both crush RAID 5 in both speed and safety.
cat /proc/mdstat
In following result, md2 shows [UU_], one drive in the raid is not functioning properly, it is recovering. This can be caused by hardware problem.
[[email protected] ~]# cat /proc/mdstat Personalities : [raid1] [raid6] [raid5] [raid4] md0 : active raid1 sdc1[2] sda1[0] sdb1[1] 6291444 blocks super 1.0 [3/3] [UUU] md1 : active raid1 sdc2[2] sda2[0] sdb2[1] 524276 blocks super 1.0 [3/3] [UUU] md2 : active raid5 sdc3[3] sda3[0] sdb3[1] 2916638720 blocks super 1.0 level 5, 512k chunk, algorithm 2 [3/2] [UU_] [============>........] recovery = 60.6% (884434512/1458319360) finish=209.6min speed=45630K/sec unused devices: <none> [[email protected] ~]#
[[email protected] ~]# mdadm --detail /dev/md2 /dev/md2: Version : 1.0 Creation Time : Wed Mar 20 04:00:24 2013 Raid Level : raid5 Array Size : 2916638720 (2781.52 GiB 2986.64 GB) Used Dev Size : 1458319360 (1390.76 GiB 1493.32 GB) Raid Devices : 3 Total Devices : 3 Persistence : Superblock is persistent Update Time : Wed Mar 20 11:04:23 2013 State : clean, degraded, recovering Active Devices : 2 Working Devices : 3 Failed Devices : 0 Spare Devices : 1 Layout : left-symmetric Chunk Size : 512K Rebuild Status : 60% complete Name : rescue:2 UUID : 7aaf9fb2:2699e2be:05a74e1f:96dd453d Events : 4130 Number Major Minor RaidDevice State 0 8 3 0 active sync /dev/sda3 1 8 19 1 active sync /dev/sdb3 3 8 35 2 spare rebuilding /dev/sdc3 [[email protected] ~]#
RAID 10
- Hardware RAID 10 is where you don't see down time for a single failed drive while maintaining speed. Don't use raid 5 it is slow.
- For high disk I/O and database server RAID 10 is must for both performance and safety.
- I do web hosting and every machine I have runs Raid-10 with 250GB SATA Drives. I have lost a drive on almost every machine I have and some machines more than one. With the SL boxes they are hot-swap so there is no down time to replace the bad drive. I primarily use it for i/o though but the fact it means less down time is great as well.
- You need 4 drives for RAID 10
- RAID - 10 has 4x read and 2xwrite speed of normal drive (+ SAS / SCSI has higher rpm). So you get good throughput.
Source: https://forums.softlayer.com/showthread.php?t=3840 (customers only)
