[Lustre-discuss] Lustre v2.1 RHEL 6.1 build does not work

[Jon Zhu]

Got it. Right now the client is running the same patched Redhat 6.1 xen
kernel as MGS and OSS node, I guess that's not the **patchless** client you
mentioned.

Does patchless client means client running on un-patched kernel? Is there a
way to build such client from the latest v2.1 code base? Or I can use
official 2.0 release client to work with v2.1 MDS/OSS nodes?

Thanks for the quick replies!

-Jon.


On Fri, Jun 24, 2011 at 5:50 PM, Oleg Drokin <green at whamcloud.com> wrote:

> Yes, in case of a patchless client I expect you'd get something like what
> you see.
> I expect that if you cd into the dir after the test is started it might
> work, though.
>
> On Jun 24, 2011, at 5:47 PM, Jon Zhu wrote:
>
> > Do you mean if this directory is continuously being deleted and
> re-created during the test so that's why the other client cannot list file
> inside it? I guess so, after I stopped the fileop test program I can get
> into the directory and there is nothing in it.
> >
> > Thanks,
> > -Jon.
> >
> >
> >
> > On Fri, Jun 24, 2011 at 5:11 PM, Oleg Drokin <green at whamcloud.com>
> wrote:
> > Did it delete the dir before creating the files, but after the other
> client did chdir?
> >
> > On Jun 24, 2011, at 5:09 PM, Jon Zhu wrote:
> >
> > > Agree with you. But here comes the real problem, when I'm running a
> program called fileop to benchmark the Lustre file system from one client,
> the other client failed to list file in the directory where the fileop
> generates files on the mounted Lustre volume. Is this a known issue?
> > >
> > > [root at test1 fileop_L1_0]# ls
> > > ls: cannot open directory .: Stale NFS file handle
> > >
> > >
> > > Thanks,
> > > -Jon.
> > >
> > >
> > >
> > >
> > > On Fri, Jun 24, 2011 at 4:06 PM, Oleg Drokin <green at whamcloud.com>
> wrote:
> > > Hello!
> > >
> > >
> > > On Jun 24, 2011, at 4:00 PM, Jon Zhu wrote:
> > >
> > > > Yes this is a Redhat 6.1 xen kernel, it works with this warning
> message, but well, I found under
> /var/spool/abrt/kerneloops-1308854953-1114-9 directory, there is a file
> called backtrace:
> > > > You are right kernel is still running but, this kerneloops-xxx
> directory name makes me think it's a  crash, maybe it's recoverable. Any
> idea?
> > >
> > > It's not a crash, just a warning.
> > >
> > > Bye,
> > >    Oleg
> > > --
> > > Oleg Drokin
> > > Senior Software Engineer
> > > Whamcloud, Inc.
> > >
> > >
> >
> > --
> > Oleg Drokin
> > Senior Software Engineer
> > Whamcloud, Inc.
> >
> >
>
> --
> Oleg Drokin
> Senior Software Engineer
> Whamcloud, Inc.
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.lustre.org/pipermail/lustre-discuss-lustre.org/attachments/20110624/e9e66ca7/attachment.htm>