1 |
On Sun, 2005-08-14 at 14:22 -0700, Robin H. Johnson wrote: |
2 |
> multiple instances on a 128-node x86 cluster (1/node). There are one or |
3 |
> two things I'd like to change in Catalyst, as they currently make it |
4 |
> slightly dangerous for your system. If for some reason an ebuild goes |
5 |
> wrong, the bind mounts allow it access into your core system, namely |
6 |
> /dev and $DISTDIR and it really hurts to have a bunch of /dev erased on |
7 |
> a box that you don't have quick physical access to (I toasted a few |
8 |
> cluster nodes due to a bug in some scripts this way, but luckily I could |
9 |
> force a remote reinstall). |
10 |
|
11 |
I accept patches. *grin* |
12 |
|
13 |
By the way, I am working to get catalyst running on OSX, so version 2.0 |
14 |
will definite suit your needs when it is released. |
15 |
|
16 |
-- |
17 |
Chris Gianelloni |
18 |
Release Engineering - Strategic Lead/QA Manager |
19 |
Games - Developer |
20 |
Gentoo Linux |