| 1 |
On Fri, Aug 05, 2011 at 09:57:08AM -0400, Rich Freeman wrote: |
| 2 |
> In any case, as long as a solution exists for md+lvm+luks+/usr before |
| 3 |
> we start breaking more stuff than is already broken, then we should be |
| 4 |
> fine. Having more than one optional solution is fine. While I don't |
| 5 |
> think that gentoo needs to be another Ubuntu, having reasonable |
| 6 |
> out-of-the-box support for one of the major desktop environments |
| 7 |
> running on md+lvm+luks seems pretty useful. |
| 8 |
|
| 9 |
My proposed minimal initramfs is an alternative to dracut for users with |
| 10 |
trivial systems. For _ALL_ of the other cases, genkernel/dracut should |
| 11 |
be used. Patches to add any of that other support to the minimal |
| 12 |
initramfs will be rejected. |
| 13 |
|
| 14 |
Definition of a trivial system: |
| 15 |
/, /usr, /var are contained on or within device nodes available in |
| 16 |
devtmpfs at the time the kernel invokes the initramfs. |
| 17 |
|
| 18 |
Examples of non-trivial systems: |
| 19 |
- {/, /usr, /var} on any of LVM, LUKS, NFS, iSCSI |
| 20 |
- {/, /usr, /var} on MD that is not auto-assembled by the kernel [1]. |
| 21 |
- etc. |
| 22 |
|
| 23 |
The initramfs would be rebuilt extremely infrequently, since it doesn't |
| 24 |
contain any kernel modules, and reads the actual configuration from the |
| 25 |
real root fs. |
| 26 |
|
| 27 |
[1]: Upstream MD kernel folk have said everybody wanting to do MD |
| 28 |
detection should migrate to initramfs for detection, as there at lot of |
| 29 |
corner cases in detection where userspace is needed to make a sane |
| 30 |
decision. Also if you use non-msdos partition tables, or large MDs. |
| 31 |
|
| 32 |
-- |
| 33 |
Robin Hugh Johnson |
| 34 |
Gentoo Linux: Developer, Trustee & Infrastructure Lead |
| 35 |
E-Mail : robbat2@g.o |
| 36 |
GnuPG FP : 11AC BA4F 4778 E3F6 E4ED F38E B27B 944E 3488 4E85 |
Archives