1 |
Thank you. I really did make initramfs my self before, but since we |
2 |
have genkernel, I did not make it myself for a long time. And other |
3 |
reason I want to use genkernel rather than do it myself is that I do |
4 |
not have physical access to that server, I control it through SSH. So, |
5 |
if I failed...I need to make a trip ;(. Anyway, I think it is a good |
6 |
time to learn how to make initramfs again ;). |
7 |
|
8 |
|
9 |
On Sun, Jun 21, 2009 at 10:59 PM, Duncan<1i5t5.duncan@×××.net> wrote: |
10 |
> David Shen <davidshen84@××××××××××.com> posted |
11 |
> 53e35fd50906201903j565753edre2148ff57729466d@××××××××××.com, excerpted |
12 |
> below, on Sun, 21 Jun 2009 10:03:51 +0800: |
13 |
> |
14 |
>> According to gentoo bug #120236, genkernel will not directly support |
15 |
>> build xen kernel. Does this means I will have to apply the patch in the |
16 |
>> bug whenever I want to build my xen kernel with genkernel? My system is |
17 |
>> build on LVM, so I need genkernel to build the start up script to active |
18 |
>> LVM. Is there a better way to build xen kernel with genkernel? |
19 |
> |
20 |
> That bit about LVM isn't correct. Think about it. LVM isn't a Gentoo |
21 |
> invention, while genkernel is a Gentoo app. What do you think Red Hat, |
22 |
> Debian and others do to support / on LVM? They aren't going to run |
23 |
> genkernel. |
24 |
> |
25 |
> While I specifically chose not to put / (or my backup /) on lvm here |
26 |
> precisely to avoid having to generate an initrd or initramfs as I wanted |
27 |
> to keep the setup simple, building an initramfs/initrd isn't Gentoo or |
28 |
> genkernel specific. There's instructions out there much as there are |
29 |
> for / on md/kernel RAID and LVM. If you're working at a sufficiently |
30 |
> advanced enough level to need both xen and LVM, and to contemplate doing |
31 |
> the xen kernel patches, then I'd suggest you're advanced enough to learn |
32 |
> how to manage the kernel, including an initramfs and/or an initrd, |
33 |
> without genkernel. |
34 |
> |
35 |
> Meanwhile, patching the kernel is no big deal. I created my own scripts |
36 |
> here to automate my own kernel building, after doing it manually for some |
37 |
> time. They apply any patches they find in my patch dir as part of the |
38 |
> process. I'm now running direct upstream git repository based kernels, |
39 |
> occasionally applying patches that haven't yet made it to mainline, but |
40 |
> while the scripts do deal with patching, as I chose a system simple |
41 |
> enough not to have to worry about an initrd/initramfs, I haven't dealt |
42 |
> with that manually and thus haven't expanded the scripts to deal with it |
43 |
> either. |
44 |
> |
45 |
> I'd suggest simply biting the bullet. Just as you evidently did in |
46 |
> learning how to setup xen and lvm in the first place, and I did with GRUB |
47 |
> (I was using LILO until recently), LVM and md/kernel-RAID among other |
48 |
> things fairly recently and the kernel itself some years ago here, I'd set |
49 |
> aside a bit of time to go thru the documentation, digesting it in enough |
50 |
> depth not only well enough to follow some step-by-step, but until I |
51 |
> groked the basic process well enough to do decent troubleshooting of any |
52 |
> problems that may come up, as necessary. That way, when the automated |
53 |
> tools like genkernel fail because you're now out in territory beyond what |
54 |
> they're designed to handle, you know the process well enough to deal. |
55 |
> |
56 |
> Just "IMHO. YMMV." =:^) |
57 |
> |
58 |
> -- |
59 |
> Duncan - List replies preferred. No HTML msgs. |
60 |
> "Every nonfree program has a lord, a master -- |
61 |
> and if you use the program, he is your master." Richard Stallman |
62 |
> |
63 |
> |
64 |
> |
65 |
|
66 |
|
67 |
|
68 |
-- |
69 |
Best Regards, |
70 |
David Shen |