[Lustre-discuss] Fwd: lustre 1.6.3 prod - client kernel panic
Matt
mattmoran76 at gmail.com
Thu Nov 8 04:51:10 PST 2007
Cheers Balagopal! I was in the process of building my own kernel, which
looks to have worked - I got the rpms in the build directory.
But I followed your commands to enable me to build it under the Centos5 with
the lustre kernel source installed. Working like a charm now!
How did you see if the interface was dropping frames?
On 07/11/2007, Balagopal Pillai <pillai at mathstat.dal.ca> wrote:
>
> Hi,
>
> Now i remember what i did to build the bnx2 sources. Here is the
> part from Lustre Manual.
>
> "
>
> The kernel-source and lustre-source packages are provided in case you
> need to build external kernel modules or use additional network types.
> They are not required to run Lustre.
>
> Once you have your Lustre source tree you can build Lustre by running
> the sequence of commands given below.
>
> $ cd <path to kernel tree>
>
> $ cp /boot/config-'uname -r' .config
>
> $ make oldconfig || make menuconfig
>
> # For 2.6 kernels
>
> $ make include/asm
>
> $ make include/linux/version.h
>
> $ make SUBDIRS=scripts
>
> "
>
> This is it!!
>
>
> Regards
> Balagopal
>
>
>
>
>
> Matt wrote:
> > I think you guys are spot on, this client is running on a Dell 1950.
> >
> > I'm in the process of rebuilding the box with the Centos 5 kernel
> > along with the source so I can patch it with lustre to build my own
> > kernel and obviously build the bnx2 driver.
> >
> > I would obviously rather just use the lustre rpm's like I have been
> > doing, but was unable to build the bnx2 driver against the lustre
> > kernel source. I'm guessing because it's not complete.
> >
> > If anyone knows any different then please say so.
> >
> > Cheers,
> >
> > Matt
> >
> > On 07/11/2007, *Balagopal Pillai* <pillai at mathstat.dal.ca
> > <mailto:pillai at mathstat.dal.ca>> wrote:
> >
> > Hi,
> >
> > Looks like the broadcom driver issue to me too. I had
> the
> > similar problem with kernel panic
> > for the Lustre kernel on Dell pe1950 with Centos 4.5 and updating
> the
> > bnx2 driver solved the issue.
> > Also the interface has the tendency to drop frames. Increasing the
> rx
> > ring parameters with ethtool
> > fixes that problem.
> >
> >
> > Regards
> > Balagopal
> >
> > Bernd Schubert wrote:
> > > Hi Matt,
> > >
> > > On Wednesday 07 November 2007 12:50:32 Matt wrote:
> > >
> > >> Hi folks,
> > >>
> > >> Built servers with Centos 5, and installed the lustre rhel5_x86
> > rpms
> > >> successfully, modified grub and booted in to the new kernel.
> > >>
> > >> Configured a mgs/mdt two OSTs and a client.
> > >>
> > >> Now when running iozone from the client I always receive a
> > kernel panic in
> > >> the lines of:
> > >>
> > >> Call trace: <IRQ> [<ffffffff88149e4b>] :
> > bnx2:bnx2_start_xmt+0x49/0x4d8
> > >> Code: 49 8b 85 e8 00 00 00
> > >> RIP [<ffffffff88146e91>] : bnx2:bnx2_poll+0xf7/0xb75 RSP <......>
> > >> CR2: 0.....
> > >> <0> Kernel Panic - not syncing : Fatal Exception
> > >>
> > >> I have attached a screenshot.
> > >>
> > >> When I bounce the client and remount I can access the lustre fs
> > fine, seems
> > >> to be purely a client issue.
> > >>
> > >> Any ideas?
> > >>
> > >
> > > looks very much like a bug in the broadcom nx2 driver. I would
> > send it to
> > > linux-netdev. And as a first try I would check whats happens
> > when you disable
> > > MSI for this driver.
> > >
> > > Cheers,
> > > Bernd
> > >
> > >
> > >
> > >
> >
> > _______________________________________________
> > Lustre-discuss mailing list
> > Lustre-discuss at clusterfs.com <mailto:Lustre-discuss at clusterfs.com>
> > https://mail.clusterfs.com/mailman/listinfo/lustre-discuss
> >
> >
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.lustre.org/pipermail/lustre-discuss-lustre.org/attachments/20071108/12652393/attachment.htm>
More information about the lustre-discuss
mailing list