1 |
On Thursday 17 November 2005 14.59, Eric Thibodeau wrote: |
2 |
> I would have scanned the mailling list for this but never found the |
3 |
> search engine for it... |
4 |
> |
5 |
> As the title says, I would love to see a Gentoo clustering solution |
6 |
> based on the www.clustermatic.org/www.ltsp.org approach (basically, |
7 |
> PXE booted OS). In fact, work on this would open up Gentoo to both |
8 |
> clustering and LTSP usage. Most of the work that needs to be done, |
9 |
> if I am not mistaken, is to create a Gentoo environment which is |
10 |
> NFS bootable as done with LTSP. We could then "easily" manage the |
11 |
> tolls and utilities available on the slave nodes by using portage. |
12 |
> Obviously, there is also some kernel tweaking involved but we can |
13 |
> start from some previous work done in that aera in all cases. |
14 |
> |
15 |
> I am actually asked to build a Gentoo based cluster next semester |
16 |
> and I would definately like to build it to be as flexible and |
17 |
> scalable as possible. I believe that the above approach fits these |
18 |
> requirements quite well. So I'm opened to any suggestions for this |
19 |
> project. Obviously, I'll document this and attempt to make the |
20 |
> procedure as accessible to anyone as possible. |
21 |
> |
22 |
> Future work would probably be to seamlessly integrate OpenPBS into |
23 |
> such a PXE-able environment to enable it to reboot/configure nodes |
24 |
> as required for given tasks/profiles. |
25 |
> |
26 |
> Thanks, |
27 |
> |
28 |
> Eric Thibodeau |
29 |
|
30 |
|
31 |
|
32 |
Hello folks |
33 |
|
34 |
This is actually quite doable with gentoo, and not too hard I might |
35 |
add. I'm surely no gentoo/linux-wiz, and I have banged together |
36 |
something like this. I have built several different clusters by the |
37 |
setup since 2002 and it seems to work ok. |
38 |
|
39 |
SSI: single system image. All servers and nodes run off the shared |
40 |
root image, no problem with local installs and duplicating binaries. |
41 |
|
42 |
Diskless boot: all nodes can boot diskless over pxe => dhcp => tftp => |
43 |
nfs |
44 |
|
45 |
Root over nfs: just one file system tree, nothing on diskless nodes. |
46 |
Swap, and local tmp storage on nodes that have disks. |
47 |
|
48 |
Openmosix, mpi, pvm, gridengine, custom batch queues etc for the |
49 |
clustering. The apps I write usually use fork and forget over |
50 |
OpenMosix. It's just soooo simple. |
51 |
|
52 |
And since openmosix helps with load balancing you can use the nodes as |
53 |
regular workstations as well. I wouldn't recommend using the servers |
54 |
as workstations for stability reasons, even if it is possible. My own |
55 |
constant pre-alpha test cluster runs from whatever I happend to have |
56 |
available, and is home-based on my workstation. But my production |
57 |
clusters have dedicated servers. |
58 |
|
59 |
A nice side effect is that most if not all of the system is somewhat |
60 |
hot-swappable. I can even swap the servers on a running cluster, with |
61 |
some restrictions. Yes I know it sounds crazy, but I've done it |
62 |
successfully several times. |
63 |
|
64 |
No duplicate system trees and binaries, just tweak with |
65 |
selective /etc/init.d scripts and such. |
66 |
|
67 |
A few init.d scripts has to be slightly changed from the gentoo |
68 |
originals, along with /sbin/rc and /sbin/functions.sh. |
69 |
The nodes can have different config files. But that is a very minor |
70 |
overhead. |
71 |
|
72 |
|
73 |
I have posted on this topic before, but there hasn't been much |
74 |
interest. If you want to try it then contact me and I'll see if I can |
75 |
whip up a quick terse description. Then we can flesh it out as we go. |
76 |
But I don't have time to actually write something nice for quite a |
77 |
while, unfortunately. |
78 |
|
79 |
Harebrafolk |
80 |
Jimmy |
81 |
-- |
82 |
gentoo-cluster@g.o mailing list |