[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-devel] Trying to unmap invalid handle! pending_idx: @ drivers/net/xen-netback/netback.c:998 causes kernel panic/reboot



On Mon, Jul 14, 2014 at 12:53:49PM +0200, Armin Zentai wrote:
[...]
> >>h14z4mzbvfrrhb was a name of a VIF. This VIF belongs to a Windows Server
> >>2008 R2 X64 virtual machine. We had 6 random reboots until now, all of the
> >>VIFs are belonged to the same operating system, but different virtual
> >>machines. So only Windows Server 2008 R2 X64 system's virtual interfaces
> >>caused the crashes, these systems has been provisioned from different
> >>installs or templates. The GPLPV driver's versions are also different.
> >>
> >
> >Unfortunately I don't have Windows server 2008 R2. :-(
> >
> >This bug is in guest TX path. What's the workload of your guest? Is
> >there any pattern of its traffic?
> 
> It's not relevant, some of them uses one core at nearly 100%, some of them
> had 1-2% CPU and 5-10 mbps of networking and/or IO. I've tried to test the
> CPU with CPU burn, prime95, tried to stress the network with SYN flood, IIS
> stress testing with apache ab, stressing the throughput and bandwidth, but
> these attempts did not caused a reboot.
> 

FYI I think the only thing that matters is the packets sent from the
guest, stressing CPU won't help. I could be wrong, but the tests you
mentioned seem to be sending packets to the guest. That's guest RX path
which is not very relevant to this bug. I think the stress test should
run in the guest, i.e. sending packets from guest to outside world. That
might help reproduce this bug. 

Wei.

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
http://lists.xen.org/xen-devel


 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.