On Mon, Jul 17, 2017 at 14:06:46 +0200, Marc Haber wrote:
> 3: veth0@tun0-vpnssh0: <BROADCAST,MULTICAST,M-DOWN> mtu 1500 qdisc noop state
> DOWN mode DEFAULT group default qlen 1000
> link/ether 42:0e:d1:a9:40:93 brd ff:ff:ff:ff:ff:ff
> 4: tun0-vpnssh0@veth0: <BROADCAST,MULTICAST,M-DOWN> mtu 1500 qdisc noop state
> DOWN mode DEFAULT group default qlen 1000
> link/ether 16:c0:48:f5:f1:6a brd ff:ff:ff:ff:ff:ff
I don't doubt that your system is doing this. That doesn't make it
right.
If I were you I would want to know why this is, when the kernel is
explicitly asked to create device names tun0-vpnssh0 and tun0-vpnssh1.
I can only reproduce what your system is doing with the following:
ip link add dev foo type veth peer
I suppose in this case the "peer" keyword tells the kernel that the peer
device will be fully specified, but the dev or name options are missing,
so it is named the default veth0.
If the "peer" keyword is not given, the kernel derives the names of both
devices from the same string. Your kernel appears to be breaking that
contract.
I can only hypothesize at this point
- maybe the kernel is misreading the info sent by iproute2
- maybe some udev/systemd device rename is happening
--
mike
