1 |
"John R. Dunning" writes: |
2 |
> From: Bryan Green <bgreen@××××××××.gov> |
3 |
> |
4 |
> It is encouraging to hear that you are willing to base a product on Lustre |
5 |
> 1.6. |
6 |
> |
7 |
> There are problems either way, but based on my experience, I believe 1.6 is a |
8 |
> better choice, at least for the kind of situation I'm expecting to see. |
9 |
> That's based partly on the fact that in my testing I've seen a pretty small |
10 |
> quotient of out-and-out bugs (though there are a couple which are pretty |
11 |
> annoying) and partly on the fact that configuration and management-wise, 1.6 |
12 |
> is way easier to deal with. Part of what I expect will be happening in |
13 |
> deployments is to be building lustrefs's on the fly, under control of some |
14 |
> kind of configurator thingie. For that kind of task, 1.4 would be much more |
15 |
> difficult to deal with. |
16 |
> |
17 |
|
18 |
>From my limited experience with 1.6, and even more limited experience with 1.4, I |
19 |
wholeheartedly agree with your assessment. Version 1.4 looks like a real headache |
20 |
to configure. By comparison, 'mount -t lustre' pretty much characterizes the |
21 |
simplicity of 1.6. |
22 |
|
23 |
> |
24 |
> Are you by any chance willing to share some of your knowledge about |
25 |
> installing Lustre on Gentoo with others? :) |
26 |
> |
27 |
> Sure. |
28 |
> |
29 |
> Are you worrying about the kernel patching and other software installation |
30 |
> issues, or about how to set up the fs itself once you've got the software |
31 |
> together? |
32 |
|
33 |
Kernel patching. For software installation, the lustre ebuild that was put on |
34 |
this list recently seemed to do the trick for me, and setup was pretty easy. |
35 |
|
36 |
I was able to patch the kernel, but the server was somewhat unstable. Actually, |
37 |
my memory is hazy. I used the 'lustre-sources' ebuild, which effectively packaged |
38 |
up the patches. It was a 2.6.15 kernel. I also tried to make a custom kernel for |
39 |
lustre 1.4, but ultimately hit too many roadblocks. I did learn a bit about how |
40 |
to use 'quilt' though. |
41 |
|
42 |
> |
43 |
> Very briefly, the kernel-patching issue is an ongoing headache. Lustre |
44 |
> patches vfs in non-trivial ways. Unfortunately, everybody else does too. It |
45 |
> becomes a fairly ugly patch-merging problem. If you want, I can detail the |
46 |
> process I've settled on for coming up with a kernel patchset, but you won't |
47 |
> like it. There are similar issues around ldiskfs and other bits, but they're |
48 |
> simpler, at least by comparison. |
49 |
|
50 |
I'd be interested in some of the details - off-list if that is more appropriate, |
51 |
though it might be of interest to others on the list as well. Once you download a |
52 |
1.6 beta, how do you produce a kernel for Gentoo? Do you patch a gentoo-sources |
53 |
kernel, a vanilla-sources kernel, or something else? The ideal would perhaps be |
54 |
to have a 'lustre-sources' ebuild in the gentoo-science overlay. :) |
55 |
|
56 |
> |
57 |
> Perhaps I could make |
58 |
> self-support an option, if it looked like it would be reliable. |
59 |
> |
60 |
> Well, obviously, you should test the bejeezus out of your configuration before |
61 |
> you declare open season on it. So far I haven't found reason to believe |
62 |
> lustre is substantially worse than any of the other open-source software |
63 |
> packages which are used in production situations. I think that constitutes a |
64 |
> qualified "yes" :-} |
65 |
|
66 |
Are you considering getting support from CFS at some point? Sorry, you don't have |
67 |
to answer if that is a sensitive question. But part of this thread has been the |
68 |
topic of encouraging CFS to support Gentoo. Interestingly, my colleague, who is |
69 |
in charge of installing Lustre (1.4) on our test system, is talking to CFS about |
70 |
supporting a vanilla kernel configuration. The reason? We can't make the system |
71 |
stable with a SLES kernel. It was stable for a long time with Gentoo. Now they |
72 |
seem to have gotten it stable with SLES plus a vanilla 2.6.19 kernel (which of |
73 |
course does not have the Lustre patches). So they want Suse to provide a newer |
74 |
SLES kernel with the Lustre patches, and CFS to support that configuration. |
75 |
|
76 |
-bryan |
77 |
|
78 |
-- |
79 |
gentoo-cluster@g.o mailing list |