I'm seeing similar. But it is "rare" it does happen.
I saw DHCP/request traffic between two VM's..
It is on a 1:1 virtual ethernet wire...
The time I saw this, is
1) I had "iface eth0 dhcp" in interfaces. The VM booted,
I then "ifdown eth0<enter>", the DHCP "release" gets displayed.
2) I edit eth0 in interfaces to
->iface eth0 inet manual , to just have
"up ip link set eth0 up"
"down ip link set down"
3)
ifup eth0
tcpdump -e -i eth0 -n<enter>
(-n to prevent dns/hostname resolving)
I noticed dhcp traffic containing the "macaddress of eth0 from the
running machine". So this is DHCP traffic FROM the right machine.
There's another thing.. and I'm filing another bug
There's port 138 broadcast traffic when I don't have samba
installed.(this is on a "veth" 1:1 wire connection)
^ This is more ridiculous. And this other bug is "definitely" much more
easily verifiable and happens all the time.
So feel free to attack that bugreport,.. the unwanted dhcp request
traffic over here is actually pretty rare but I've noticed again and again.
If you want to re-open this bugreport it is up to you.
...When I wanted to re-confirm my setup on another hypervisor system,
the "delay" with with openvswitch only complicated things. Back to the
original hypervisor setup, I've been able to confirm the dhcp problem it
really does happen.
It seems to occur soon after an eth0 interface setup change, from a
"dhcp" to "static" or "manual", .. But it would make more than 1 dhcp
request. This tends to happen on an interface which does not have an ip
address bound to it.
The reason why I don't have an ip address on a NIC interface is because
I intend to use it with a nids/snort sensor setup and so there should be
no traffic generated from this interface..
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
