RAID 5 and disk capacity
You mention that fewer than four disks (three) on RAID-5 isn't worth it. Why?
Software/Hardware used:
Software/Hardware used:
ASKED:
July 7, 2009 3:00 PM
UPDATED:
July 8, 2009 1:54 PM
Answer Wiki:
Seeming that Raid5 needs 3 disks at a minimum you do not have any room or drives for failure. It has nothing to switch over to and if any one of the three drives fails the entire array will fail. If you have four or five drives you can potentially loose one or two drives without the array being totally offline - slow maybe but your data would still be available. I would use a raid0 configuration instead/.
To see all answers submitted to the Answer Wiki: View Answer History.
If redundancy in the array is a requirement, then 3 disks is the bare minimum, but as KarlG indicates, there’s no room for error. The array will fail, but if memory serves, you won’t lose data; you’ll have to physically swap out the failed disk before you can recover.
If performance is the driving force, then RAID0 or a JBOD array does just fine, but a loss of a single drive will corrupt the filesystem created on the array.
Also – don’t try anything other than RAID0 or RAID1 at the OS level. The performance hit on the CPU is horrendous. If you want a protected array with decent performance, get a hardware RAID controller that supports RAID5 (RAID3 if you’re feeling adventurous) and as many disks as you can feed it, keeping in mind that the usable size of a RAID5 array is total aggregate capacity of all physical disks, minus one disk’s size for the dedicated parity store.