Ipv6 link local address left on old openvpn interface after restarting openvpn

[stefb]

Hi,

I run into an issue with openvpn tap in client mode, : a link local address is added to the tap interface, and when i restart the client it dies with this error :

Jul 3 19:05:20	openvpn[1007]: TUN/TAP device ovpnc11 exists previously, keep at program end
Jul 3 19:05:20	openvpn[1007]: TUN/TAP device /dev/tap11 opened
Jul 3 19:05:20	openvpn[1007]: do_ifconfig, tt->ipv6=0, tt->did_ifconfig_ipv6_setup=0
Jul 3 19:05:20	openvpn[1007]: /sbin/ifconfig ovpnc11 10.8.63.2 netmask 255.255.255.0 mtu 1500 up
Jul 3 19:05:20	openvpn[1007]: FreeBSD ifconfig failed: external program exited with error status: 1
Jul 3 19:05:20	openvpn[1007]: Exiting due to fatal error

and even if the tunnel is down i see the old tap interface still exiting and configured only with a link local address

ovpnc11: flags=8051 <up,pointopoint,running,multicast>metric 0 mtu 1500
	options=80000 <linkstate>inet6 fe80::290:bff:fe23:7630%ovpnc11 prefixlen 64 scopeid 0x12</linkstate></up,pointopoint,running,multicast> 

Why wouldn't  openvpn destroy the interface when it's killed / stopped.

This is a serious problem as i can't restart any of my tap tunnels. I don't see the same issue when using tun devices.

Additionnal info : I'm running quagga/ospf on the tap interfaces

Edit: the link local address is the same as the link-local address of my first ethernet port, not a random generated address by openvpn

[stefb]

Ok i think i found what caused this : the set tos of packet option was selected, it seems that if i remove that option, restart of openvpn doesn't fail… quite odd it doesn't fail at the first start.