1 |
From: "Daniel van Ham Colchete" <daniel.colchete@×××××.com> |
2 |
Date: Mon, 4 Dec 2006 21:55:12 -0200 |
3 |
[...] |
4 |
Lustre uses a lot of kernel features that if not enabled will |
5 |
cause the kernel to crash. |
6 |
[...] |
7 |
I don't think that's true. |
8 |
|
9 |
I've been running lustre on assorted kernels, mostly under gentoo dists, for |
10 |
some months, and found that once you get past the installation issues, it's |
11 |
pretty trouble free. |
12 |
|
13 |
Now, note the caveats there: The installation issues are non-trivial, mostly |
14 |
because lustre is very intrusive into the vfs layer. This causes no end of |
15 |
headaches integrating with various other peoples' kernel patches, due to |
16 |
collisions with the other peoples' patches to vfs. That statement is as true |
17 |
of gentoo as it is of any kernel other than the few for which they supply |
18 |
canned patchsets. But I've never seen anything in there that contitutes using |
19 |
a kernel feature which causes the kernel to crash if not enabled. The closest |
20 |
thing I've seen to that is if you muff the patch merging and end up with an |
21 |
inconsistent patchset, that generally leads to a crash :-} |
22 |
|
23 |
Lustre 1.6 (at least the client end) doesn't even really *require* all those |
24 |
kernel patches, ie they do support the idea of a patchless client. The issue |
25 |
is that lustre changes the logic involved in various kinds of fs operations, |
26 |
including anything related to lookups, so as to short-circuit much of the work |
27 |
involved when it figures out that it can do so. Running the client without |
28 |
the patches will work, but it won't give you the performance that you'd get |
29 |
with the patches. So odds are anybody who's interested in running lustre in |
30 |
the first place probably wants the patches too. |
31 |
|
32 |
Lustre also is not restricted to precompiled kernels, their build script |
33 |
contains patchsets for things other than their recommended redhat and suse |
34 |
ones. We routinely compile it up for all kinds of experimental kernels with |
35 |
no trouble. Again, once you've gotten over the hurdle of getting the kernel |
36 |
patches integrated, the rest of it behaves reasonably well. The reason cfs |
37 |
advocates the small number of kernels they do is because they know what a pain |
38 |
it is to draw outside the lines, and they try to steer people away from that. |
39 |
|
40 |
We at sicortex are planning on rolling out a gentoo-based cluster that depends |
41 |
heavily on lustre, so we've spent a fair bit of time banging on it. I'm |
42 |
pretty sure we understand it at this point. We'll know for sure soon :-} |
43 |
-- |
44 |
gentoo-cluster@g.o mailing list |