I am having issues with RA on my RB450G, and this is what I have discovered, If I do not slave any ports on the 450G RA works fine and all devices on my LAN get IPv6 addresses, however if I slave together ports 3,4 and 5 to port 2, this breaks RA because the RB gets confused as to which MAC address belongs to which port so the link local address is incorrect for the ether port, it seems to be a random process when you reboot the RB as to which MAC address is tied to which port so the dynamic fe80 addresses are incorrect. This is very easy to reproduce, it is currently working at the moment after 5-6 reboots of the RB because the correct MAC/Port config has occured. I have emailed support also and sent a supout file. This happens with V 4.3 4.4 4.5 and 4.6. I am not the only one who is having this problem, I know of 3 others.
I have this exact problem as well. When switch ports are slaved on my RB750, RADVD tries to advertise on them. In the logs the error is: “radvd debug sendmsg failed on private (fe80:20c:42ff:fe55:b518): Invalid argument” When I disable the slave ports RADVD works correctly, but only after a router reboot.
Same problem here with RB750G and Router OS 5.0 beta1.
RADVD does not send any packets if ports are slaved.
Good point : IPv6 over PPPoE is working now.
How can this bug be present since so long ?
Seems like IPv6 is a toy for geeks actually. When we’ll need it really to clients side nothing will work correctly…
We are waiting a serious effort from Mikrotik for IPv6. Seems like version 5.0 is a minor update, i can’t understand why Mikrotik slowed down so much their developpment timing, did they loose their master programmers or project leader ?
Same problem here with a 750G on 4.11 - will be opening another bug report with support as it seems the ones from other people in this thread didn’t go anywhere. Didn’t try any betas as this router is part of my home network, and the wife acceptance factor for the Internet randomly not working is fairly low.
Tried rebooting a few times but the router reliably picks the address of the slaved port every time and advertisements fail.
Why would it even generate a link local address for a port that is part of a switch group? Only the master port should have an IPv6 address at all.