Hi,

I have a Centos7 host with libvirt, attempting to boot a Centos7 guest. I am attempting to do the install over ipv6 (because why not).

I have the following network configuration.

Cisco 3550 gi0/1 <-> eno1 (Centos7) host.

gi0/1 is configured as:

switchport trunk encapsulation dot1q switchport trunk allowed vlan 1-17 switchport mode trunk

Mainly in question is vlan11 as this is the network I am trying to work with.

The switch have vlan11 configured with STP:

# sh spa v 11 VLAN0011 Spanning tree enabled protocol rstp Root ID Priority 12299 Address 000c.30d1.9700 This bridge is the root Hello Time 2 sec Max Age 10 sec Forward Delay 5 sec

Bridge ID Priority 12299 (priority 12288 sys-id-ext 11) Address 000c.30d1.9700 Hello Time 2 sec Max Age 10 sec Forward Delay 5 sec Aging Time 300

Interface Role Sts Cost Prio.Nbr Type ------------------- ---- --- --------- -------- -------------------------------- Gi0/1 Desg FWD 4 128.25 P2p Gi0/2 Desg FWD 4 128.26 P2p

Notice that the switch is the root of the STP.

The Centos7 host network is configured as:

DEVICE="eno1.11" VLAN="yes" BOOTPROTO="none" IPV6INIT="no" IPV6_AUTOCONF="no" NM_CONTROLLED="no" ONBOOT="yes" TYPE="Ethernet" BRIDGE=br11

With the corresponding bridge as:

DEVICE=br11 ONBOOT=yes BOOTPROTO=none TYPE=Bridge IPV4_FAILURE_FATAL=no IPV6_FAILURE_FATAL=no NM_CONTROLLED=no IPV6INIT=yes IPV6_AUTOCONF=yes IPV6_ROUTER=no STP=off

The output of ip a for these interfaces is as such:

5: eno1.11@eno1: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue master br11 state UP link/ether c0:3f:d5:61:c8:ec brd ff:ff:ff:ff:ff:ff inet6 fe80::c23f:d5ff:fe61:c8ec/64 scope link valid_lft forever preferred_lft forever 6: br11: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue state UP link/ether c0:3f:d5:61:c8:ec brd ff:ff:ff:ff:ff:ff inet6 fe80::c23f:d5ff:fe61:c8ec/64 scope link valid_lft forever preferred_lft forever

brctl show br11

bridge name bridge id STP enabled interfaces br11 8000.c03fd561c8ec yes eno1.11 vnet1

vnet1 being the interface of the connected vm.

When the VM is booted, the following error is received.

[ OK ] Reached target Paths. [ OK ] Reached target Basic System. [ 33.874572] IPv6: eth0: IPv6 duplicate address 2001:db8:0:11:5054:ff:fedd:a617 detected!

Any address assigned to the interface immediately triggers a dadfailed warning such as:

2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000 link/ether 52:54:00:dd:a6:17 brd ff:ff:ff:ff:ff:ff inet6 2001:db8:0:11::200/64 scope global tentative dadfailed valid_lft forever preferred_lft forever inet6 fe80::5054:ff:fedd:a617/64 scope link valid_lft forever preferred_lft forever

Assigning an ipv4 address, traffic is correctly able to flow over the bridge via the switch to other hosts.

I am suspicious that the issue lays somewhere in ND being reflected from the cisco in some capacity perhaps?

A fedora 20 system connected to gi0/2, with identical network configuration and switch does not suffer this issue when booting guests.

This has left me quite stumped as to what I am doing wrong on the Centos7 host, or if there is another issue.

And help would be greatly appreciated in helping solve this issue.