1 |
* Joakim Tjernlund <joakim.tjernlund@×××××××××.se> [120126 05:38]: |
2 |
[..] |
3 |
> |
4 |
> When you update the RO layer you will be back to a single copy, the previous |
5 |
> release which is only in the overlay is gone, right? |
6 |
|
7 |
That's up to you. You can manage your updates however you want. For |
8 |
example, you can have your RO layer loop mounted out of a release image |
9 |
and have two (or more) different RW areas depending on which image |
10 |
you're running. That can allow you to fallback from an update fairly |
11 |
easily. |
12 |
|
13 |
> |
14 |
> It also implies 2 different SW update methods, one for updating the overlay |
15 |
> copy and one for updating the RO layer. |
16 |
|
17 |
Well, you probably will have an initramfs to set up the aufs. As part |
18 |
of the initramfs you can do whatever you have to for the update. If you |
19 |
have things on the RW layer that should be carried over you can leave |
20 |
them if you mount the same RW filesystem with the new RO layer or you |
21 |
can copy them to a new RW filesystem if you want them to move forwards |
22 |
but still want a separate RW filesystem for each image to allow easy |
23 |
fallback. It's really up to how you want to manage it and how much RW |
24 |
you need to go forwards with the update (and where it is on the |
25 |
filesystem too.) |
26 |
|
27 |
[..] |
28 |
> AFAIK, only FSes which are designed for flash(JFFS2, UBIFS, YAFFS etc. ) |
29 |
> are safe w.r.t power cuts. |
30 |
|
31 |
Yes, that's certainly been my experience. |
32 |
|
33 |
[..] |
34 |
> |
35 |
> We are not using NAND flash yet but our next product will. I do have the |
36 |
> impression that any block emulating device such as SSD are unreliable |
37 |
> w.r.t power cuts. I would love to be proven wrong though :) |
38 |
> |
39 |
> Jocke |
40 |
|
41 |
I've used NOR flash with aufs and a RO loop mounted root filesystem from |
42 |
within a release image and UBIFS RW filesystem. |
43 |
|
44 |
It worked well. |
45 |
|
46 |
I've used NOR with JFFS2 before as well and had some issues. Those may |
47 |
have been bugs in the JFFS2 implementation I used though. |
48 |
|
49 |
Todd |