Re: oops with dual xeon 2.8ghz 4gb ram +smp, software raid, lvm, and xfs

[Jakob Oestergaard]

On Tue, Nov 23, 2004 at 09:37:44AM -0600, Phil Dier wrote:
...
> I'm building this system with stability and flexibility foremost in
> mind. Am I foolish in using all of these technologies with a new-ish
> version of 2.6? Is there a particular version that would be better
> suited for my application? Any other suggestions you (or anyone else
> on the list) could give regarding stability would be greatly appreciated.

If you'll be exporting via. NFS, it seems that there are still problems
with XFS+NFS.

With SMP, what I see is that sometimes a directory might decide that
it's a file - but I can't delete it, becuase it isn't 'empty' (it's
still somehow a directory).  Waiting a day or two, the system will
change its mind back to letting the directory be a directory. Sometimes
modes will be fscked up as well - a regular file can change owner, or it
can change modes from '-rw-rw---' to '?---------'.    Weird stuff, no
way to reproduce it reliably.

With UP, I know someone who's seeing stale handles reported by the NFS
server. The only known workaround is to stat the directories in question
on the *server* side - a little bash with 'while true; sleep 5; ls -l
/directory; do' will do the trick.

All of what I describe here are production environments - so it sucks to
have that kind of problems.  Some of it can be reproduced (the stale
handle errors), and some of it can't.

I guess the good news would be, that I don't know of any problems with
XFS+LVM+MD if you do not export the FS via. NFS    :)

That is, if you run 2.6.9.  Any earlier kernel will b0rk your XFS under
load.

-- 

 / jakob

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/