IPv6 PPPoE Mikrotik server and Windows client

Hi, I’m just getting into IPv6 and trying to make PPPoE work with Windows client.
So far I got Windows to connect and to get prefix, but I have problem getting DNS server to Windows. PPPoE server works on ether1_hosts and in ND DNS is advertised, but Windows doesn’t get it. (when I use SLAAC everything works)

 ipv6 nd print 
Flags: X - disabled, I - invalid, * - default 
 0  * interface=all ra-interval=3m20s-10m ra-delay=3s mtu=unspecified 
      reachable-time=unspecified retransmit-interval=unspecified 
      ra-lifetime=30m hop-limit=unspecified advertise-mac-address=yes 
      advertise-dns=yes managed-address-configuration=no 
      other-configuration=yes 

 1    interface=ether1_hosts ra-interval=3m20s-10m ra-delay=3s mtu=unspecified 
      reachable-time=unspecified retransmit-interval=unspecified 
      ra-lifetime=30m hop-limit=64 advertise-mac-address=yes advertise-dns=yes 
      managed-address-configuration=no other-configuration=yes 

 2    interface=<pppoe-test> ra-interval=3m20s-10m ra-delay=3s mtu=unspecified 
      reachable-time=unspecified retransmit-interval=unspecified 
      ra-lifetime=30m hop-limit=64 advertise-mac-address=yes advertise-dns=yes 
      managed-address-configuration=no other-configuration=yes

Also, pool for PPPoE clients is not obtained via DHCP-Client, I inserted it manually and I didn’t make route to it, because I expected when client connects, Mikrotik will see new connected route and RIPng will distribute it to all other routers but that doesn’t happen even thou I have RIPng active on all interfaces. Why? I can bypass this “issue” by inserting static route for ipv6-pool.

You would need to configure RIPng to redistribute static routes.

You should use OSPF unless you know your network is going to stay small and simple.

Thanks for the tip, I kinda expected it to be connected route rather than static.
I’m all for OSPF, but right now main admin is against it because “we only have single gateway and we don’t need it”. So in IPv4 we use RIPv2, but when we begin implementing IPv6 into production I will try use OSPF. Thing is, we mostly have single route to any host only with few exceptions where multiple routes are available, so even thou network is medium sized ~100 subnets it’s all relatively simple for routing.

RIP broadcasts the entire routing table (well, the entire RIP table to be fair) every interval. RIP would use a single 56k hop to avoid taking two 10-gigabit hops. (Mikrotik’s default OSPF configuration behaves the same, until you put costs on interfaces).

Any time I’m teaching network to a junior admin, I tell them there’s only two things to remember about RIP: “RIP sucks.”

RIP is fine for a relatively stable, small network, but other than that, use a link-state protocol like OSPF.