Hello World
I am fighting some weird IPv6 RA issue.
Situation:
- Bridge with two Ethernet Interfaces, both connected to two different units of a stacked Cisco Switch.
- RSTP enabled. (one is root port, the other one is alternte port)
- IGMP Snooping disabled (breaks IPv6 ND)
- Bridge has an IPv6 Address bound with 'advertise' enabled.
- IPv6 ND is configured for 'all' interfaces.
On a Linux box, reception of IPv6 RA can be verified:
tcpdump -vvvv -ttt -e -i eth0 'icmp6 and ('ip6[40] = 134' or 'ip6[40] = 133')'
- Add VLAN Interface, VLAN2
- Configure the Bride Interface to tag VLAN2 on both Ethernet Interfaces.
- Bind IPv6 Address on VLAN 2 interface, 'advertise' enabled.
- Make sure Cisco Interfaces are in Trunk mode and allow VLAN 2 on both ports to the Mikrotik and to the Linux Client used for testing.
- Set-up Vlan2 on the Linux client, interface eth0.2
Use above tcpdump to verify if IPv6 RA are being received. Sometimes they are not.
Sometimes disabling end re-enabling the IPv6 IP Address on the Mikrotik on the VLAN2 interface solves the issue. Sometimes 'togging' IPv6 ND solves the issue. Sometimes a Reboot of the Mikrotik solved the issue.
Today I found the next weird solution: Check STP status on both Ethernet Interfaces on Bridge. 'Disable' the root port one, the alternate becomes forwarding. IPv6 RA are being sent out on VLAN 2.
Re-Enable previous root-port, the roles witch back. IPv6 RA are now being sent over the port from which they were not being sent before.
I have been comparing the setting of both interfaces and bridge interfaces side-by-side, they are configured equal, the counterpart on the cisco is configured equal.
I don't get why IPv6 RA sometimes just stop being sent.
I am aware of the STP / PVSTP difference between Cisco and Mikrotik, but basically every 'trunk port' has the same VLAN assigned to them, so this should not be the issue.
Has anyone a cloue what could cause the issue or is observing the same?
-Benoît-