[ovirt-users] [OT] how to analyze/avoid dropped rx packets

[Gianluca Cecchi]

Hello,
on some Oracle Linux 7 VMs I see that dropped RX packets continue to
increase.
The eth0 interface is virtio for all of them.
When running tcpdump on eth0 the counter stop increasing and so it means
tcpdump does capture them

I found some references about these kind of rx dropped frames with CentOS 7
too and other distros with new kernels.

eg:
https://www.netiq.com/support/kb/doc.php?id=7007165
and
*https://serverfault.com/questions/528290/ifconfig-eth0-rx-dropped-packets
<https://serverfault.com/questions/528290/ifconfig-eth0-rx-dropped-packets>*

In my case it doesn't seem that the nature of the dropped packages is STP,
but IP6

On my VMs I have disabled ipv6

I see that in about 100 seconds I have 180 dropped frames occured, and when
running tcpdump for the same amount of time I have in it about 180 DHCPv6
packets, so I presume it is it the responsible
Any way to avoid this, at VM or hypervisor level?

Thanks in advance,
Gianluca
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.ovirt.org/pipermail/users/attachments/20171110/52c687cb/attachment.html>