1 |
Hi Charles |
2 |
|
3 |
I've been looking for time to answer this more fully than the quick |
4 |
oneshot mail I send off earlier. |
5 |
We run 600 servers on gentoo and started with a single server originally. |
6 |
|
7 |
So yes it's definitely doable :) |
8 |
|
9 |
I'll try and answer as much as possible, any questions feel free to mail. |
10 |
|
11 |
Charles Duffy wrote: |
12 |
> I'm looking at replacing SuSE SLES9 with Gentoo for an enterprise |
13 |
> application (for reasons of flexibility and licensing) (no, we don't |
14 |
> have an enterprise application budget -- just the reliability |
15 |
> requirements; yaaay, startups!). We're looking to be able to deploy and |
16 |
> manage hundreds of geographically distributed servers. |
17 |
See above, what is your planned initial deployment ? Are you starting |
18 |
with a hundred or more servers or are you starting with just a couple ? |
19 |
> We have a QA department available to vet each configuration before it is |
20 |
> deployed to the field. We have infrastructure for tracking the progress |
21 |
> of code in svn from creation though QA to deployment; I'm anticipating |
22 |
> tracking a local overlay (containing all packages we use), make.conf, |
23 |
> /etc/portage/*, etc. through this system, autobuilding system images |
24 |
> (either to run virtualized or on real hardware) from the contents of |
25 |
> svn, building binary packages and deploying them to real hardware. |
26 |
Having a QA department to offload this work to is certainly a bonus :) |
27 |
> I'm interested in best practices, suggested tools, and/or 3rd party |
28 |
> experiences in this regard. |
29 |
> |
30 |
> Some particular questions which come to mind: |
31 |
> - Should I be using a custom profile or a standard profile with |
32 |
> overrides through make.conf, /etc/portage/* and the like? |
33 |
AFAIK you should be able to set all required stuff through overrides. |
34 |
The point is to keep in mind the benefits of using gentoo and not try |
35 |
and work against the system. |
36 |
Several people discussed running gentoo servers in produktion without a |
37 |
build toolchain (gcc etc.) |
38 |
I have no comments to offer on how desirable this is, but if this is a |
39 |
goal/requirement for you I'd strongly suggest using a binary distro. |
40 |
Gentoo shines as a ultra-configurable source-based distro, running it |
41 |
without a build toolchain and / or portage tree is certainly possible. |
42 |
It would however take away much of the advantages of using gentoo, so |
43 |
why not switch to something else in that case. |
44 |
|
45 |
Removing the portage tree has always been a weird question to me, nobody |
46 |
discusses removing the rpm-package database, why are people so keen on |
47 |
removing the portage tree ? |
48 |
It takes roughly 550Mb of space which is quite a lot, but hardly a |
49 |
killing requirement given todays diskspace and hardware. |
50 |
The linux kernel source tree is roughly in the same sizing category. |
51 |
> - What's the Right Way to create new system images ready to be loaded |
52 |
> onto a hard drive or run through a virtual machine? gentoo-buildhoster |
53 |
> looks interesting. I've seen Catalyst mentioned as a way to create |
54 |
> stage3 images, but what documentation I've been able to find doesn't |
55 |
> seem very much targeted for my use cases. |
56 |
I would recommend catalyst-2. Although documentation is lacking, it |
57 |
isn't that hard to setup. |
58 |
You're probably looking for the stage4 target if you want to build |
59 |
system images. |
60 |
Rolling out gentoo on such a large scale, you need a repeatable system |
61 |
image build environment. |
62 |
|
63 |
The bonus of catalyst is that you automatically get a binary package |
64 |
server in the process of generating your images. |
65 |
Catalyst can be told to use a binary package cache, by carefully setting |
66 |
up your catalyst environment you can easily reuse that as source for |
67 |
your binpkg server. |
68 |
I'm not familiar with gentoo-buildhoster, but since it's webpage [1] |
69 |
lists it as no longer maintained that would be a no-go area for me. |
70 |
|
71 |
We combined catalyst with a pxe based boot environment, the quickstart |
72 |
installer [2] and puppet[3]. |
73 |
It allows us to provision a server within 30 minutes. That's 30 minutes |
74 |
from connecting the hardware to a switch and active in production. |
75 |
This requires very little manual intervention, which we consider to be a |
76 |
good thing (tm) |
77 |
And yes, that's concurrent, we believe it to be capable of roughly 30 |
78 |
servers concurrent setup and that appears to be a pxe limitation. |
79 |
|
80 |
If at all possible, try to build your deployment system thus that you |
81 |
can always easily wipe a server and reinstall. |
82 |
We didn't originally and are refactoring to allow it now. |
83 |
> - Any experiences with puppet? With out ratio of servers to staff, |
84 |
> automating configuration and administration is a priority. (We already |
85 |
> have an internal tool written with automating the server configuration |
86 |
> process in mind; it has some functionality puppet doesn't, and puppet |
87 |
> has functionality it doesn't; in theory, I'd like to extend puppet until |
88 |
> our internal tool becomes unnecessary, though I'll need to understand |
89 |
> puppet much better before I can think too hard about that). |
90 |
We are using puppet extensively. It works, although it's still rough |
91 |
around the edges which is as expected from such a young project. |
92 |
The gentoo provider for puppet is in it's infancy. It works, but |
93 |
definitly needs work as well. |
94 |
|
95 |
Apart from that puppet is a very versatile and powerful tool. |
96 |
And most importantly it has a very active community of people around it. |
97 |
They are actively exchanging recipes for server configuration, which is |
98 |
useful in itself but becomes extremely useful when combined with |
99 |
the new module organization in puppet. |
100 |
|
101 |
Many problems you will face in deploying and configuring such an amount |
102 |
of servers will have been solved in whole or in part by someone in the |
103 |
puppet community. |
104 |
|
105 |
I would be curious what functionality is missing from puppet right now |
106 |
in your opinion ? |
107 |
> - Have any of 'yall been in the 100s-of-servers situation with |
108 |
> comparable requirements and come up with a different approach to |
109 |
> managing it? How did things work out? |
110 |
We've grown very very fast and have tried different methods along the way. |
111 |
To be honest, we're still in the process of moving most of the |
112 |
serverpark under puppet control (nearly 50% done) |
113 |
And I actually do not expect to find a single set of tools that cope |
114 |
with all the issues that you face when deploying such an amount of servers. |
115 |
Your situation might be different because you are starting with a single |
116 |
app. |
117 |
|
118 |
One vital thing that is missing from the picture is an inventory database. |
119 |
You need some sort of queryable database that stores servers, location, |
120 |
networking info, function, server-identifyable serial of some sort, |
121 |
hardware classes, deployment status etc. |
122 |
Without that you're basically lost. We use a homebrew mysql based system |
123 |
with both a cli- and a webinterface. |
124 |
|
125 |
Apart from that, you'll need lots of infrastructure: |
126 |
* logservers |
127 |
* monitoring |
128 |
* statistics gathering |
129 |
* backups |
130 |
* scripts repository |
131 |
* version controlled configs |
132 |
* bug / issue tracking |
133 |
* firewalling |
134 |
* loadbalancing |
135 |
* network monitoring and configuration system |
136 |
* etc. etc. etc. |
137 |
> Thank you! |
138 |
You're welcome, I would like to see more people use gentoo in |
139 |
large-scale environments and am actively looking for possibilities to |
140 |
exchange experiences. |
141 |
Feel free to contact me if you have any questions. I also hang out at a |
142 |
couple of irc-channels [4] with nickname Innocenti |
143 |
|
144 |
Out of pure curiosity what is your staff to server ratio ? |
145 |
|
146 |
Grtz Ramon |
147 |
|
148 |
Senior System Administrator Hyves.nl |
149 |
|
150 |
[1]http://badpenguins.com/source/ |
151 |
[2]http://dev.gentoo.org/~agaffney/quickstart.php |
152 |
[3]http://puppet.reductivelabs.com/ |
153 |
[4]#puppet, #gentoo-server, #gentoo-cluster, #gentoo-amd64 |
154 |
-- |
155 |
gentoo-server@g.o mailing list |