1 |
Before someone jumps in and says "that's not true"... |
2 |
|
3 |
...That's not true... |
4 |
|
5 |
hehe... From what I see, onesis is a kit of wrapper scripts which seem to |
6 |
automate most of the procedure on my wiki. The problem with wrapper scripts |
7 |
is that they have to be maintained and, when it comes to automating system |
8 |
configuration, they tend to restrict that only _THEY_ perform modifications |
9 |
on the config files. Another example of this is LTSP, you can either choose |
10 |
to use their ltspadmin interface to configure DHPC, pxeboot, NFS and all the |
11 |
other stuff or you do it manually. In my case, I do it manually and the |
12 |
scripts don't recognize that I have done the job. |
13 |
|
14 |
so... |
15 |
Pros of doing it manually: |
16 |
- You have total control |
17 |
- You _have_ to know what you are doing |
18 |
- you're not limited to the path imposed by the tool (this applies to both |
19 |
distributions and "admin" scripts) |
20 |
|
21 |
Cons: |
22 |
- You have to control everything |
23 |
- You _have_ to know what you are doing |
24 |
- There is a steep learning curve in most cases |
25 |
- Sysadmins (to be read as MCSE people) dUn't like it |
26 |
|
27 |
This said, Josh, have you tried it yourself?...comments?...feedback ;) |
28 |
|
29 |
Eric |
30 |
|
31 |
Le Mardi 11 Avril 2006 15:39, Simon-Nicolas Roth a écrit : |
32 |
> No, |
33 |
> |
34 |
> I prefer the unionfs way. To me it is more simple to have one / for all |
35 |
nodes |
36 |
> and another / for the server. |
37 |
> |
38 |
> Simon |
39 |
> |
40 |
> Le Tuesday 11 April 2006 16:13, Josh England a écrit : |
41 |
> > Have you tried using oneSIS for going diskless? |
42 |
> > |
43 |
> > emerge -va onesis |
44 |
> > |
45 |
> > Check out http://onesis.org for more info. |
46 |
> > |
47 |
> > -JE |
48 |
> > |
49 |
|
50 |
-- |
51 |
Eric Thibodeau |
52 |
Neural Bucket Solutions Inc. |
53 |
T. (514) 736-1436 |
54 |
C. (514) 710-0517 |
55 |
|
56 |
-- |
57 |
gentoo-cluster@g.o mailing list |