On Thu, Oct 26, 2006 at 02:04:01PM -0600, Andreas Dilger wrote:
> On Oct 26, 2006 16:21 +0200, email@example.com wrote:
> > http://www.paul.sladen.org/vserver/archives/200506/0058.html
> > During compilation you should have gotten warnings about implicit
> > declarations of vx_rmap_pid. Take the .c files for which those
> > warnings show up, and add an #include <linux/vs_cvirt.h> to them.
> > That should fix your issues.
> It would make more sense to include this header into whatever defines
> the macros that affect the compiled files. That way no need to change
> the compiled sources at all?
not so trivial, for various reasons, but this has
been fixed in a different way in recent patches
> > ----- Original Message -----
> > From: "Andreas Dilger" <firstname.lastname@example.org>
> > To: "Jon Scottorn" <email@example.com>
> > Cc: "lustre-discuss" <firstname.lastname@example.org>
> > Sent: Thursday, October 26, 2006 12:46 AM
> > Subject: Re: [Lustre-discuss] Lustre 1.6-beta Debian package with 2.6.18
> > asuccess
> > >On Oct 25, 2006 13:48 -0600, Jon Scottorn wrote:
> > >>Do you know what I might be missing to get the patchless client
> > >>working on debian with vservers.
> > >>
> > >>I can get the patchless client working without vservers and I had
> > >>heard that someone had it working with vservers. I can build the
> > >>lustre source patchless client with the vservers patched in the
> > >>kernel but then when I go to mount lustre FS I get a bunch of
> > >>unknown symbols reported to dmesg.
> > >>
> > >>Here is a sample of a few:
> > >>
> > >>libcfs: Unknown symbol vx_rmap_pid
> > >
> > >This is caused by the vserver patch, I'd guess. Maybe it adds a
> > >macro that Lustre calls somewhere that accesses this symbol? Can't
> > >help you there.
> Cheers, Andreas
> Andreas Dilger
> Principal Software Engineer
> Cluster File Systems, Inc.
> Vserver mailing list
Vserver mailing list
Received on Thu Oct 26 23:12:23 2006