[fedora-virt] Slow virtual ping

[Andrés García]

Hello,

I am having a problem trying to setup a F16 machine as a virtual host.

For example, I have another F16 installed as a virtual guest. If I ping it
just after the host boots I get something like:

PING 192.168.0.158 (192.168.0.158) 56(84) bytes of data.
64 bytes from 192.168.0.158: icmp_seq=1 ttl=64 time=0.358 ms
64 bytes from 192.168.0.158: icmp_seq=2 ttl=64 time=0.377 ms
64 bytes from 192.168.0.158: icmp_seq=3 ttl=64 time=0.317 ms
64 bytes from 192.168.0.158: icmp_seq=4 ttl=64 time=0.336 ms
64 bytes from 192.168.0.158: icmp_seq=5 ttl=64 time=0.293 ms
64 bytes from 192.168.0.158: icmp_seq=6 ttl=64 time=0.334 ms
64 bytes from 192.168.0.158: icmp_seq=7 ttl=64 time=0.353 ms
64 bytes from 192.168.0.158: icmp_seq=8 ttl=64 time=0.350 ms
64 bytes from 192.168.0.158: icmp_seq=9 ttl=64 time=0.291 ms
[...]

Which is very reasonable, but after some time, I can't tell you
how much exactly, maybe half an hour, I get:

PING 192.168.0.158 (192.168.0.158) 56(84) bytes of data.
64 bytes from 192.168.0.158: icmp_seq=1 ttl=64 time=22.5 ms
64 bytes from 192.168.0.158: icmp_seq=2 ttl=64 time=20.8 ms
64 bytes from 192.168.0.158: icmp_seq=3 ttl=64 time=19.1 ms
64 bytes from 192.168.0.158: icmp_seq=4 ttl=64 time=17.9 ms
64 bytes from 192.168.0.158: icmp_seq=5 ttl=64 time=16.8 ms
64 bytes from 192.168.0.158: icmp_seq=6 ttl=64 time=15.2 ms
64 bytes from 192.168.0.158: icmp_seq=7 ttl=64 time=14.1 ms
64 bytes from 192.168.0.158: icmp_seq=8 ttl=64 time=13.2 ms
64 bytes from 192.168.0.158: icmp_seq=9 ttl=64 time=11.7 ms
64 bytes from 192.168.0.158: icmp_seq=10 ttl=64 time=9.76 ms
64 bytes from 192.168.0.158: icmp_seq=11 ttl=64 time=7.86 ms
64 bytes from 192.168.0.158: icmp_seq=12 ttl=64 time=6.63 ms
64 bytes from 192.168.0.158: icmp_seq=13 ttl=64 time=5.27 ms
64 bytes from 192.168.0.158: icmp_seq=14 ttl=64 time=3.79 ms
64 bytes from 192.168.0.158: icmp_seq=15 ttl=64 time=1.83 ms
64 bytes from 192.168.0.158: icmp_seq=16 ttl=64 time=99.8 ms
64 bytes from 192.168.0.158: icmp_seq=17 ttl=64 time=97.9 ms
64 bytes from 192.168.0.158: icmp_seq=18 ttl=64 time=96.8 ms
64 bytes from 192.168.0.158: icmp_seq=19 ttl=64 time=95.2 ms
64 bytes from 192.168.0.158: icmp_seq=20 ttl=64 time=94.5 ms
64 bytes from 192.168.0.158: icmp_seq=21 ttl=64 time=92.9 ms
64 bytes from 192.168.0.158: icmp_seq=22 ttl=64 time=91.9 ms
64 bytes from 192.168.0.158: icmp_seq=23 ttl=64 time=90.7 ms
64 bytes from 192.168.0.158: icmp_seq=24 ttl=64 time=88.9 ms
64 bytes from 192.168.0.158: icmp_seq=25 ttl=64 time=87.9 ms
64 bytes from 192.168.0.158: icmp_seq=26 ttl=64 time=87.7 ms
64 bytes from 192.168.0.158: icmp_seq=27 ttl=64 time=85.9 ms
64 bytes from 192.168.0.158: icmp_seq=28 ttl=64 time=84.8 ms
[...]

Now ping answers are much slower for no reason I can see and
they don't get back to 'normal' unless I reboot the host.

I have two network interfaces, the one in the motherboard for the
host to use and another one (PCI) which is bridged for the guests.

The host keeps answering pings in less than 1ms without a problem.

The guest has a virtio network interface.

During these tests the F16 guests is the only guest running, I have also
tried them with a Win7 guest and pretty much the same thing happens.

Do you have any idea what could be the problem?

Thanks,
Andres