Robert Ficcaglia
2007-Apr-07 07:53 UTC
[Xen-users] network error, even when reboot into non-Xen smp kernel
After installing xen-3.0.4_1-install-x86_32p.tgz on a CentOS 4.4, I was getting a network error: Device eth0 has different MAC address than expected I have searched for hours here and other places, tried all the recommended solutions and no luck. However, my real question is this... Even after I reboot into the previous CentOS kernel ( CentOS (2.6.9-42.0.10.ELsmp)), I get the same network error. I have tried this on 2 different boxes and it is completely reporducible. (Different network cards, too.) Strangely, I can boot into the non-smp kernel: CentOS (2.6.9-42.0.10.EL) and get networking. What is Xen doing that is persistent and causing network failure even after fallback to the original smp kernel?? (Maybe it is worth saying that the network was 100% fine before Xen install, in fact, sadly it is a completely remote server, so that is the only way to get into the box unless I want to pay for an expensive remote hands service!!) How can I undo these changes and restore networking to my smp kernel? Help please!! _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Nico Kadel-Garcia
2007-Apr-09 09:59 UTC
Re: [Xen-users] network error, even when reboot into non-Xen smp kernel
Robert Ficcaglia wrote:> After installing xen-3.0.4_1-install-x86_32p.tgz on a CentOS 4.4, I > was getting a network error: > Device eth0 has different MAC address than expectedFirst, why aren''t you using the RHEL 4 RPM? It''s true they accidentally left out the documentaiton in that, but it works fine, and has RPM kernels for both Dom0 and DomU. Second, unless you hard-code your MAC address in your Xen configuration file, it''s going to get changed randomly in your DomU''s every freaking time you reboot them. And RedHat''s network setup tools look in /etc/sysconfig/network-scripts/ifcfg-eth* to associate the HWADDR setting with specific network devices. This keeps eth0 and eth1 from getting swapped around for different kernel configurations, and helps keep the network device names consistent when adding wi-fi or USB network devices. You can also simply comment out the HWADDR lines in /etc/sysconfig/networ-kscripts/ifcfg-eth* if you have only one network device and thus aren''t worried about swapping them. Is the problem in your Dom0? Do you have multiple network devices?> I have searched for hours here and other places, tried all the > recommended solutions and no luck. However, my real question is > this... Even after I reboot into the previous CentOS kernel ( CentOS > (2.6.9-42.0.10.ELsmp )), I get the same network error. I have tried > this on 2 different boxes and it is completely reporducible. > (Different network cards, too.) > > Strangely, I can boot into the non-smp kernel: CentOS > (2.6.9-42.0.10.EL) and get networking.See above.> > What is Xen doing that is persistent and causing network failure even > after fallback to the original smp kernel?? (Maybe it is worth saying > that the network was 100% fine before Xen install, in fact, sadly it > is a completely remote server, so that is the only way to get into the > box unless I want to pay for an expensive remote hands service!!)_______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Robert Ficcaglia
2007-Apr-09 15:26 UTC
Re: [Xen-users] network error, even when reboot into non-Xen smp kernel
No reason I used the tarball other than it was first on the list :) I would be glad to remove what''s on there now and reinstall with RPMs, part of the problem is that I don''t know the steps to completely remove Xen and start clean; this is the best list I found: http://lists.xensource.com/archives/html/xen-users/2006-04/msg00617.html or download the src tarball and make uninstall (probably will be my first try) To answer your other questions, yes - the trouble is with Dom0, and yes I have multiple (2) NICs (same hw). So with 2 NICs, I presume I cannot simply comment out the HWARDDR (cause I tried that and it didn''t help); so what is the procedure for 2 NICs? On 4/9/07, Nico Kadel-Garcia <nkadel@gmail.com> wrote:> > Robert Ficcaglia wrote: > > After installing xen-3.0.4_1-install-x86_32p.tgz on a CentOS 4.4, I > > was getting a network error: > > Device eth0 has different MAC address than expected > First, why aren''t you using the RHEL 4 RPM? It''s true they accidentally > left out the documentaiton in that, but it works fine, and has RPM > kernels for both Dom0 and DomU. > > Second, unless you hard-code your MAC address in your Xen configuration > file, it''s going to get changed randomly in your DomU''s every freaking > time you reboot them. And RedHat''s network setup tools look in > /etc/sysconfig/network-scripts/ifcfg-eth* to associate the HWADDR > setting with specific network devices. This keeps eth0 and eth1 from > getting swapped around for different kernel configurations, and helps > keep the network device names consistent when adding wi-fi or USB > network devices. > > You can also simply comment out the HWADDR lines in > /etc/sysconfig/networ-kscripts/ifcfg-eth* if you have only one network > device and thus aren''t worried about swapping them. > > Is the problem in your Dom0? Do you have multiple network devices? > > > I have searched for hours here and other places, tried all the > > recommended solutions and no luck. However, my real question is > > this... Even after I reboot into the previous CentOS kernel ( CentOS > > (2.6.9-42.0.10.ELsmp )), I get the same network error. I have tried > > this on 2 different boxes and it is completely reporducible. > > (Different network cards, too.) > > > > Strangely, I can boot into the non-smp kernel: CentOS > > (2.6.9-42.0.10.EL) and get networking. > See above. > > > > > What is Xen doing that is persistent and causing network failure even > > after fallback to the original smp kernel?? (Maybe it is worth saying > > that the network was 100% fine before Xen install, in fact, sadly it > > is a completely remote server, so that is the only way to get into the > > box unless I want to pay for an expensive remote hands service!!) > >_______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Nico Kadel-Garcia
2007-Apr-09 17:29 UTC
Re: [Xen-users] network error, even when reboot into non-Xen smp kernel
Robert Ficcaglia wrote:> No reason I used the tarball other than it was first on the list :) I > would be glad to remove what''s on there now and reinstall with RPMs, > part of the problem is that I don''t know the steps to completely > remove Xen and start clean; this is the best list I found: > http://lists.xensource.com/archives/html/xen-users/2006-04/msg00617.html > or download the src tarball and make uninstall (probably will be my > first try) > > To answer your other questions, yes - the trouble is with Dom0, and > yes I have multiple (2) NICs (same hw). So with 2 NICs, I presume I > cannot simply comment out the HWARDDR (cause I tried that and it > didn''t help); so what is the procedure for 2 NICs?Well, a "tar tzf" of the tarball would be a good start to know what to remove. Are the NIC''s identical models to each other? Do you have available the right modules for it, as listed in /etc/modprobe.conf and listed by lsmod? _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Luciano Miguel Ferreira Rocha
2007-Apr-10 09:09 UTC
Re: [Xen-users] network error, even when reboot into non-Xen smp kernel
On Mon, Apr 09, 2007 at 08:26:56AM -0700, Robert Ficcaglia wrote:> To answer your other questions, yes - the trouble is with Dom0, and yes I > have multiple (2) NICs (same hw). So with 2 NICs, I presume I cannot simply > comment out the HWARDDR (cause I tried that and it didn''t help); so what is > the procedure for 2 NICs?Could you please run the following command in both cases? (It works, non-smp centos kernel; it doesn''t, dom0.) /sbin/ifconfig -a And send us the output? -- lfr 0/0 _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users