1 |
On Thursday 27 Jul 2017 09:48:43 symack wrote: |
2 |
> There must be an easy way to do this. Something like download the latest |
3 |
> portage and source package. Untar on live system and rebuild! |
4 |
> That would be so amazing if possible. |
5 |
|
6 |
There are a number of ways you could approach this. |
7 |
|
8 |
1. Offline Update |
9 |
|
10 |
Emerge within a chroot on an offline system having set the same CFLAGS with |
11 |
your production box, build binary packages, copy them over and run: |
12 |
|
13 |
emerge -uaNDkv @system |
14 |
|
15 |
Run dispatch, gcc-config, @represerved-rebuild, depclean and the like and |
16 |
assuming all is good reboot in the new kernel and repeat for world. |
17 |
|
18 |
This will ensure minimal downtime. |
19 |
|
20 |
2. Online Update |
21 |
|
22 |
If downtime can be tolerated and your system is fast, you can back up /etc, |
23 |
/var/lib/portage/world and your database(s), take the system offline and |
24 |
reinstall with the latest stage 3 filesystem and portage from a LiveCD. This |
25 |
will take considerably longer. |
26 |
|
27 |
3. Hybrid |
28 |
|
29 |
Personally, I would opt for a combo of 2 and 1. In other words, I would re- |
30 |
install offline using the configuration and databases of the live system, then |
31 |
test the new build to make sure all is working as expected. Finally, take the |
32 |
prod system offline, and tar over the whole filesystem from the recently |
33 |
rebuilt mirror. There may be some minor tweaks needed for fstab, hosts files |
34 |
between the two, but otherwise you will end up with a fresh build on your |
35 |
production box with relatively small amount of downtime. You will also have |
36 |
an offline/cold standby system to manage emergencies and emerges with. |
37 |
|
38 |
-- |
39 |
Regards, |
40 |
Mick |