1 |
On Fri, Nov 03, 2006 at 08:47:52AM +0000, Steve Long wrote: |
2 |
> As to why I don't just do it myself, I think it's a bit silly to duplicate the |
3 |
> compile that devs do anyway. |
4 |
My compiles as a dev are of very minimal use to anybody except me. |
5 |
There are too many things that are specific to my systems. |
6 |
|
7 |
> I appreciate that many will be against this idea, but I'd still like to |
8 |
> discuss it: a binary repository for gentoo. |
9 |
> |
10 |
> Yes, I know gentoo is a meta-distro. And that there isn't loads of bandwidth. |
11 |
> That's easily got round. The main problem I see is USE flags (devs already |
12 |
> compile with standard C-flags right?) but I was thinking about standardising |
13 |
> for 2 or 3 types of network- SOHO, medium and large enterprise (eg for LDAP |
14 |
> etc) would solve most cases. We can always tag pkgs with USE flags. |
15 |
> |
16 |
> If gentoo is still serious about enterprise adoption, it needs a binary repo |
17 |
> (so we can avoid system breakage) which would of course be a little bit |
18 |
> behind. I'd be happy to contribute time, as I'm sure many other users would. |
19 |
From all of the large Gentoo deployments I've done (one of which |
20 |
exceeded 200 machines), you're approaching this the wrong way. |
21 |
|
22 |
1. Consider where each enterprise needs customization: USE-flags, CFLAGs |
23 |
This might be for example an LDAP or Kerberos-based shop, so they would |
24 |
have their stuff built with those, or a hardened or selinux place. Or |
25 |
SASL, or Java or any of a hundred different variables. This was after |
26 |
all the point of USE flags in the first place. |
27 |
2. From point one, it is clear that each enterprise needs to be able to |
28 |
customize. They also need binaries. So we need a solution that combines |
29 |
the two. |
30 |
3. The solution is for each enterprise to have their own tinderbox / |
31 |
build-machine. Tinderboxing is supported under catalyst, and I believe |
32 |
there is at least one other tinderbox implementation around. |
33 |
4. (Assuming catalyst, as it's the only tinderbox I'm familiar with) The |
34 |
enterprise defines a specfile that describes each of their unique |
35 |
environments, and feeds these to tinderbox. Tinderbox generates sets of |
36 |
binpkgs for each environment, which the enterprise then deploys. |
37 |
|
38 |
The above plan works perfectly - I use it in my enterprise deployments. |
39 |
To use the example of my largest deployment that I mentioned above, my |
40 |
specfiles were for the following: |
41 |
- cluster nodes [128 machines] |
42 |
- cluster master [1 machine] |
43 |
- web servers [~60 machines] |
44 |
- ldap servers [2 machines] |
45 |
- dedicated file serving (network homedirs) [4 machines] |
46 |
- infrastructure management [3 machines] |
47 |
- desktops [~60 machines] |
48 |
|
49 |
The build-box role was actually undertaken by one of the infrastructure |
50 |
management machines, since it doesn't need a dedicated machine. |
51 |
|
52 |
-- |
53 |
Robin Hugh Johnson |
54 |
E-Mail : robbat2@g.o |
55 |
GnuPG FP : 11AC BA4F 4778 E3F6 E4ED F38E B27B 944E 3488 4E85 |