1 |
Hello list, |
2 |
|
3 |
I'm setting my i5 workstation up to be an emerge host to a new Atom N550 box |
4 |
on the LAN. I've created a 32-bit chroot at /mnt/atom on the i5 box, written |
5 |
an init.d script to mount the Atom's whole file system on /mnt/atom/target |
6 |
over NFS, and set an alias |
7 |
tmerj='emerge --root=/target --jobs=5 --keep-going' |
8 |
in /mnt/atom/etc/profile.d/profile_aliases.sh. |
9 |
|
10 |
The Atom box has had nfs-utils installed at an early stage so that the |
11 |
workstation can do all the heavy work. Its /etc/exports file says |
12 |
/ 192.168.2.6(sync,rw,no_root_squash,no_subtree_check |
13 |
|
14 |
This appears to be working, but for an odd wrinkle (and it's too early to |
15 |
declare success of the whole operation). Some packages, such as MySQL, when |
16 |
I've installed them this way have also been installed in parallel on the |
17 |
host machine. At least, that appeared to be happening because for several |
18 |
packages I had one job emerge " to /target/" and another emerge without that |
19 |
qualification. Yet when I cat /mnt/atom/var/lib/portage/world, it's empty. |
20 |
|
21 |
How should I interpret this? Have I missed something in the setup? I'm also |
22 |
unsure of the effect of /proc and /sys being nfs-exported from the Atom, so I |
23 |
think I'll make the export a bit more selective (when gentoo-sources have |
24 |
finished being written to disk - it's been about 20 minutes so far). |
25 |
|
26 |
Incidentally, one thing I've noticed is that the final stage of installing |
27 |
gentoo-sources, the writing to disk of all the files, takes many minutes |
28 |
because of the slow 2.5" disk in the Atom box. The rest of the operation, so |
29 |
far, is pleasingly quick - much better than building natively on the Atom. |
30 |
|
31 |
-- |
32 |
Rgds |
33 |
Peter |