I have not been able to upgrade my systems to XCP 0.5 yet, but I have some additional information that might help someone get to the cause of this error.
My setup is quite simple. I have one Debian Lenny VM acting as a firewall/router (FW) between the WAN and the LAN - traffic from the LAN VMs to the ISP passes through this machine. I'm using Corosync/Pacemaker to provide HA - it switches the global IP to an identical FW VM on a different physical XCP host in case of failure. This kicks in perfectly when the kernel panic occurs.
It seems to me like the error only occurs when the system has been running for a while. On a Windows VM on the LAN I go to
www.speedtest.net and perform a bandwidth test. This causes the FW to panic before the first test can finish, but when I restart the FW and do more tests, the panic no longer happens. During normal operation, which involves very little traffic it must be said, I've never seen the error.
I can't find the call trace for the kernel panic in the VM logs, and the free online OCR tools seem to translate my screenshot wrongly. So I'm providing the call trace as text here:
xennet_poll+0x94d/0x9ef [xen_netfront]
tick_program_event+0x1f/0x23
net_rx_action+0x9c/0x177
__do_softirq+0x45/0x53
irq_exit+0x35/0x69
do_IRQ+0x52/0x63
xen_evtchn_do_upcall+0x7/0xc
_stext+0x3a7/0x1000
xen_safe_halt+0xd/0x17
xen_idle+0x0/0x3a
xen_idle+0x2b/0x3a
xen_idle+0xb0/0xd0
============================
Code: 00 00 00 00 c7 43 04 00 00 00 00 89 10 89 42 04 8b 44 24 04 89 43 14 0f b6 54 24 0c 89 14 24 89 d6 83 bc 97 c4 08 00 00 00 74 04 <0f> 0b eb fe 89 9c 97 c4 08 00 00 8d 87 c4 0c 00 00 e8 03 22 9f
xennet_alloc_rx_buffers+0x176/0x26f [xen_netfront] SS:ESP 0069:c0393e34
Kernel panic - not syncing: Fatal exception in interrupt
Any help is much appreciated.
Thanks!
Trygve
On Thu, Jul 15, 2010 at 9:01 AM, Trygve Sanne Hardersen
<trygve@xxxxxxxxxxxxx> wrote:
I did some more testing today with a fully patched Debian Lenny VM, but unfortunately I got the same error after a short time.
I'm planning an upgrade to XCP 0.5 soon, does anybody know if that could help?
Thanks
Trygve
On Mon, Jul 5, 2010 at 7:25 PM, Trygve Sanne Hardersen
<trygve@xxxxxxxxxxxxx> wrote:
Hi,
I'm running XCP 0.1.1 and have a virtual debian-lenny machine configured as a router/firewall using iptables. I've had the same kernel panic twice now during testing of the bandwidth from a machine behind the router.
Attached is the screen dump from XVP Viewer. Any idea what is wrong? I'm no kernel expert but it seems to me like the crash happens somewhere in the Xen code.
I've now performed a dist-upgrade on the virtual machine and hopefully that will help. Will let you know if it happens again.
Thanks!
Trygve
--
HypoBytes Ltd.
Trygve Sanne Hardersen
Akersveien 24F
0177 Oslo
Norway
hypobytes.com+47 40 55 30 25
