1 |
On Mon, 21 Nov 2005, Robin H. Johnson wrote: |
2 |
|
3 |
> On Mon, Nov 21, 2005 at 12:45:53PM -0500, Eric Thibodeau wrote: |
4 |
> > PXE-boot even for the Master?...so where do the images reside... |
5 |
> The PXE env resides on an install server, that's used for all of the |
6 |
> machines in the network (not just the cluster). (It serves up SUSE's |
7 |
> AutoYAST for our desktops, and a few different GLI profiles for our |
8 |
> other servers). |
9 |
Ayee... okay, so your setup is not _strictly_ Beowulf but more of a |
10 |
Beowulfish cluster. I should have mentionned that my intention is to come |
11 |
up with a procedure/scripts/profile(s) intended for a dedicated/classic |
12 |
Beowulf cluster. |
13 |
|
14 |
> > how do you manage the slightly varying config items such as hostname |
15 |
> > and all? |
16 |
> hostname is taken from DHCP directly. dhcpcd -H. |
17 |
> The other slightly varying stuff is done with a postinstall script run |
18 |
> by the installer (the current script is under 100 lines). |
19 |
> |
20 |
> > This approach still seems a little bit time consuming since all nodes |
21 |
> > are still individual entities (not NFS roots to a single maintained |
22 |
> > image). Though granted that the nodes being all identical, emerge -K |
23 |
> > should in theory be a breeze....but it's not the case for maintaining |
24 |
> > all the config files consistent. |
25 |
> Use INSTALL_MASK=/etc after the install is completed and propagate |
26 |
> config files manually (well triggerable command connecting to a local |
27 |
> rsync server). |
28 |
> |
29 |
> We also make use of LDAP for authentication. |
30 |
|
31 |
I think I fogot to mention that the entire procedure has to fit within 15 |
32 |
pages with as little "frill" as possible. Notably, i was explicitally |
33 |
instructed _not_ to implement "complexe" authentication schemes 9such as |
34 |
LDAP+Kerberos/Windows)... |
35 |
|
36 |
> > So switching a machine's purpose/profile requires a complete |
37 |
> > re-install on the node? |
38 |
> Yes, but I'm talking about radical alterations here. |
39 |
> Eg we've got a binary app that needs the system compiled with gcc-3.3. |
40 |
> It doesn't work if glibc or a few other things are compiled with |
41 |
> gcc-3.4 or newer. |
42 |
|
43 |
This doesn't rule out using a collection NFS roots... Bare in mind that i |
44 |
am an LTSP user/implementer... so I've at leas had _some_ experience with |
45 |
the principle but haven't had to implement it from A to Z as of yet. |
46 |
|
47 |
> > You state 20 minutes for re-installing, is it |
48 |
> > a _real_ install |
49 |
> Full GLI install, extract stage, portage tree, emerge -K all needed |
50 |
> packages. If you want to make each node compile everything, it's about |
51 |
> an hour on the nodes (they are pretty speedy boxes). |
52 |
> |
53 |
> > or the dump of a "reference" root? (Pardon my |
54 |
> > ignorance of the CLI installer you are referring to... I'll read the |
55 |
> > http link you'll send me ;) ) |
56 |
> Look for GLI first of all. It's the Gentoo installer. It has a number |
57 |
> of frontend modes - GTK, Dialog, and CLI. |
58 |
> The CLI is entirely non-interactive. |
59 |
|
60 |
Hmm...will do then. I've been installing Gentoo "by hand" since 1.4 and |
61 |
never sought/required another technique... so how do you call the CLI |
62 |
gentoo Installer? Do you change the default behaviour in your DHCP |
63 |
server's config so that the PXE env. that is booted automatically triggers |
64 |
the installation of such or such profile? |
65 |
|
66 |
Thanks, |
67 |
|
68 |
Eric |
69 |
|
70 |
-- |
71 |
gentoo-cluster@g.o mailing list |