Re: [Xen-users] dom0 networking getting screwd at random periods

[Martins Lazdans <marrtins@xxxxxxxx>]

It seems that troublemaker was NICs afterall. I put some braodcom dual 
gbit into server and after that it's been 36h and no problems at all. 
Before that i was using built-int nvidia NICs (forcedeth driver). It's 
strange however why restarting bridges helped?
Todd Deshane wrote:
> On Wed, Jun 15, 2011 at 5:26 AM, Martins Lazdans <marrtins@xxxxxxxx> wrote:
> > Does anyone have any hint where to dig? Right now I wrote a script that
> > checks ping to gateway and in case of timeouts restarts bridge and
> > reattaches domUs to it. Right now bridge is stopping working like every 3
> > minutes. Can this be hardware related?
> You may want to consider setting up the bridges manually with the
> CentOS networking scripts.
>
> This is actually the default way of doing things starting with Xen 4.1
> http://wiki.xensource.com/xenwiki/HostConfiguration/Networking
>
> The standard bridging built into CentOS may end up being more stable for you.
>
> Hope that helps.
>
> Thanks,
> Todd
>
> > One guy replied to me directly having the same problem on CentOS (I guess it
> > was CentOS).
> >
> > Martins Lazdans wrote:
> > > Hello!
> > >
> > > I having a problem with dom0 (running Debian lenny, latest patches
> > > applied) networking getting stopped after random periods of time - it's been
> > > 18 hours and it's been 5 hours. dom0 got two NIC (peth0, peth1), two subnets
> > > (a, b), configured two bridges (eth0, eth1 - each subnet resides on
> > > different NIC). dom0 have each NIC configured with dedicated IP form
> > > correspongin subnets. So, eth0 serves subnet a, while eth1 is serving subnet
> > > b.
> > >
> > > In xend-config.sxp:
> > > (network-script network-bridge-wrapper) and network-bridge-wrapper is
> > > just:
> > >
> > > #!/bin/sh
> > > /etc/xen/scripts/network-bridge "$@" netdev=eth0
> > > /etc/xen/scripts/network-bridge "$@" netdev=eth1
> > >
> > > After random periods of time, eth0 goes down along with all domU
> > > networking attached to it. I'm unnable to connect or ping dom0 neither on
> > > peth0 nor peth1 nor any of domU attached to that bridge.
> > >
> > > However, domU using eth1 bridge are working just fine. I can SSH to domU
> > > and then SSH to dom0 on peth1 IP address (remember, I could not do that form
> > > "outside").
> > >
> > > After I do `/etc/xen/scripts/network-bridge stop` and then start, dom0
> > > networking comes back. I just do `brctl addif eth0 vifX.0` to attach domUs
> > > back and all works fine till next such event.
> > >
> > > Anyone had experienced such thing? I've been running this server for some
> > > three years w/o problems. This problem started around 31 of may. Maybe it
> > > has something to do with any of latest updates?
> > >
> > > There are no network spikes, no high load avg, nothing in logfiles,
> > > nothing on xm dmesg, ne domU restarts. I've been searching Google for some
> > > days now, and nothing comes up.
> > >
> > > Basicly, I've got the same issue as described here:
> > > http://lists.xensource.com/archives/html/xen-users/2008-03/msg00609.html
> > >
> > > Btw, I've got very similar issue twice with other sarver, differend
> > > hardware, different data center, CentOS 5.6, Xen 3.4.2. However, that server
> > > was using only one NIC so I'm not aware if this was excacly the same
> > > problem. I ran into this problem twice, about two months ago. However there
> > > is no errors since then which made me believe problem was solved with some
> > > OS updates? Back then I thought it's beeing hardware error as I had this
> > > server deployed only 3 months before incident.
> > >
> > > My config is below.
> > >
> > > Many thanks!
> > >
> > > # uname -a
> > > Linux doom 2.6.26-2-xen-amd64 #1 SMP Tue Jan 25 06:13:50 UTC 2011 x86_64
> > > GNU/Linux
> > >
> > > # xm info
> > > host                   : doom
> > > version                : #1 SMP Tue Jan 25 06:13:50 UTC 2011
> > > machine                : x86_64
> > > nr_cpus                : 8
> > > nr_nodes               : 1
> > > cores_per_socket       : 4
> > > threads_per_core       : 1
> > > cpu_mhz                : 2200
> > > hw_caps                :
> > > 178bf3ff:efd3fbff:00000000:00000110:00802001:00000000:000007ff
> > > total_memory           : 32767
> > > free_memory            : 2116
> > > node_to_cpu            : node0:0-7
> > > xen_major              : 3
> > > xen_minor              : 2
> > > xen_extra              : -1
> > > xen_caps               : xen-3.0-x86_64 xen-3.0-x86_32p hvm-3.0-x86_32
> > > hvm-3.0-x86_32p hvm-3.0-x86_64
> > > xen_scheduler          : credit
> > > xen_pagesize           : 4096
> > > platform_params        : virt_start=0xffff800000000000
> > > xen_changeset          : unavailable
> > > cc_compiler            : gcc version 4.3.1 (Debian 4.3.1-2)
> > > cc_compile_by          : waldi
> > > cc_compile_domain      : debian.org
> > > cc_compile_date        : Sat Jun 28 09:32:18 UTC 2008
> > > xend_config_format     : 4
> > --
> > Martins Lazdans
> >
> > _______________________________________________
> > Xen-users mailing list
> > Xen-users@xxxxxxxxxxxxxxxxxxx
> > http://lists.xensource.com/xen-users
--
Martins Lazdans

_______________________________________________
Xen-users mailing list
Xen-users@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-users