| 1 |
On 21/12/2022 20:40, Frank Steinmetzger wrote: |
| 2 |
> Yes? In a mirror setup, all member drives of a mirror have the same content |
| 3 |
> (at least in ZFS). |
| 4 |
> |
| 5 |
> Raid 10 distributes its content across several mirrors. This is the cause |
| 6 |
> for its increased performance. So when one of the mirrors (not single drive, |
| 7 |
> but a whole set of mirrored drives) fails, the pool is gone. |
| 8 |
> |
| 9 |
>> Linux will happily give you a 2-copy mirror across 3 drives - 3x6TB drives |
| 10 |
>> will give you 9TB useful storage ... |
| 11 |
|
| 12 |
> I admit, I’ve never head of that. (Though it sounds like raid-5 to me.) |
| 13 |
|
| 14 |
Raid 5 has a parity drive (or rather, raid 4 has a parity drive. Raid 5 |
| 15 |
smears parity across all disks). It does not store duplicate copies. |
| 16 |
Raid 10 has duplicate data and no parity. |
| 17 |
> |
| 18 |
Read up on linux raid-10. It is NOT raid-1+0. |
| 19 |
|
| 20 |
Drive sda sdb sdc |
| 21 |
|
| 22 |
Blocks 1 1 2 |
| 23 |
2 3 3 |
| 24 |
4 4 5 |
| 25 |
5 6 6 |
| 26 |
|
| 27 |
etc ... |
| 28 |
|
| 29 |
https://raid.wiki.kernel.org/index.php/What_is_RAID_and_why_should_you_want_it%3F |
| 30 |
|
| 31 |
(Disclaimer - I either wrote or heavily edited it.) |
| 32 |
|
| 33 |
Cheers, |
| 34 |
Wol |
Archives