| 1 |
-----BEGIN PGP SIGNED MESSAGE----- |
| 2 |
Hash: SHA1 |
| 3 |
|
| 4 |
Eric Edgar wrote: |
| 5 |
> ... the busybox discussion that you are |
| 6 |
> mentioning wont be implemented. There is too much work to get a usable |
| 7 |
> system based on that. I know for a fact that much of the regular |
| 8 |
> expression stuff will break because of that. Who knows how much else |
| 9 |
> will be broken. |
| 10 |
|
| 11 |
Yep, that's an understatement ;) found that out last night after a few |
| 12 |
more hours playing with baselayout. Sorry about the silly question ;) |
| 13 |
|
| 14 |
> For what you are doing however you could use cow files to make things |
| 15 |
> more secure. |
| 16 |
|
| 17 |
Well, really my ideal goal is to produce a system image which has _no_ |
| 18 |
persistent state whatsoever, i.e. it boots off a rom-type filesystem |
| 19 |
(squashfs), does some linking and copying before /sbin/init into a tmpfs |
| 20 |
and does a pivot_root to that (basically the same thing the livecd's |
| 21 |
initrd does, only with a directory tree on the root fs, instead of a |
| 22 |
loop-mounted image), and an init script downloads the service configs |
| 23 |
and nfs-mounts the content at startup. This way, every time updated |
| 24 |
packages are available, you just rebuild the root fs, and reboot the VM |
| 25 |
with the new image, and there's no need to copy over any of the |
| 26 |
configuration. Likewise, if you want to modify the config, you just edit |
| 27 |
the host's copy and reboot the VM, no need to change anything in the |
| 28 |
VM's root fs. |
| 29 |
|
| 30 |
> When catalyst 2 comes out you can take the stage4 target which makes it |
| 31 |
> possible to make a system that is stored in a tarball and take some |
| 32 |
> scripts and migrate it into the uml loop back files. This may be what |
| 33 |
> you are looking for. |
| 34 |
|
| 35 |
Actually that would be perfect. I've already scraped together some |
| 36 |
scripts that will take a stage3 and wrap it in a squashfs with a setup |
| 37 |
script that does the livecd-style link and copy into tmpfs before |
| 38 |
init... so if stage4 can generate the system tarball, that should do it. |
| 39 |
|
| 40 |
- -- V. Condino |
| 41 |
-----BEGIN PGP SIGNATURE----- |
| 42 |
Version: GnuPG v1.4.1 (GNU/Linux) |
| 43 |
|
| 44 |
iD8DBQFCxD6jW5g8mX2/4TYRAoMQAJ9pbSmgbG8oe4OwUOemv+EW2tCiHgCeIGOT |
| 45 |
ADeiRbQqSsipjOn1XBSKglk= |
| 46 |
=hG64 |
| 47 |
-----END PGP SIGNATURE----- |
| 48 |
-- |
| 49 |
gentoo-catalyst@g.o mailing list |
Archives