1 |
On 29/01/2019 19:01, Rich Freeman wrote: |
2 |
>> It would surely be a bug if the kernel were capable of manipulating RAIDs, but not of initialising |
3 |
>> and mounting them. |
4 |
|
5 |
> Linus would disagree with you there, and has said as much publicly. |
6 |
> He does not consider initialization to be the responsibility of kernel |
7 |
> space long-term, and prefers that this happen in user-space. |
8 |
> |
9 |
> Some of the lvm and mdadm support remains for legacy reasons, but you |
10 |
> probably won't see initialization of newer volume/etc managers |
11 |
> supported directly in the kernel. |
12 |
> |
13 |
Actually, the kernel isn't capable of manipulating raid. The reason you |
14 |
need raid 0.9 (or 1.0, actually) is that all the raid metadata is at the |
15 |
*end* of the partition, so the system boots off the file-system |
16 |
completely ignorant that it's actually a raid. The raid then gets |
17 |
assembled and when root is remounted rw, it's the raid version that's |
18 |
mounted not a single-disk filesystem. |
19 |
|
20 |
In other words, if you have sda1 and sdb1 raided together as your root, |
21 |
the system will boot off a read-only sda1 before switching to a |
22 |
read-write md1. |
23 |
|
24 |
Cheers, |
25 |
Wol |