IPv6 troubles with multiple static addresses in CHR

RouterOS General
1

Hello!
I’m trying to setup IPv6 in my VPS with CHR (v 7.15). VPS provider gave me a /64 prefix and fe80::1 as gateway. Here’s config on my side (first 4 blocks of addresses are omitted):

/ipv6 address
add address=0000:0000:0000:0000:: advertise=no interface=ether1
add address=0000:0000:0000:0000:abc:: advertise=no interface=ether1
add address=0000:0000:0000:0000:cba:: advertise=no interface=ether1
add address=0000:0000:0000:0000:cba:123:: advertise=no interface=ether1
add address=0000:0000:0000:0000:cba:abc:: advertise=no interface=ether1
/ipv6 firewall filter
add action=accept chain=input connection-state=established,related
add action=accept chain=input protocol=icmpv6
add action=accept chain=input dst-port=46380 protocol=udp
add action=accept chain=input in-interface=WG-v6 protocol=ospf
add action=accept chain=input dst-port=8291 protocol=tcp
add action=drop chain=input
add action=accept chain=forward in-interface=WG-v6 out-interface=ether1
add action=drop chain=forward
/ipv6 nd
set [ find default=yes ] advertise-dns=no disabled=yes interface=ether1 ra-interval=20s-1m
/ipv6 route
add disabled=no dst-address=::/0 gateway=fe80::1%ether1 routing-table=main suppress-hw-offload=no

The problem is that when I ping 0000:0000:0000:0000:: it seems to instead respond with the lexicographically last address (so now it’s :cba:abc:: - if I disable it, then :cba:123:: and so on). Is this a bug, or maybe the “subnet IP” is not intended to be white? They keep rebooting me to Linux and telling that everything is working on it, so I should check my configuration.

2

Is there anything I can do about this? Adding multiple IPv6 addresses works bad - some of them work flawless, some have packet loss and huge ping… Is this a thing to blame the provider or it’s a fault in my configuration?
In traceroute I can see, that previous hop before my host timeouts, is that normal for IPv6?

3

What do you need multiple addresses for? If you want to hand out ip’s to your WG peers, just route the appropriate subnets to the interface / same for container addresses.

4

See section 2.6.1 of RFC 4291 - IP Version 6 Addressing Architecture

An address where Interface ID ends with 0s is a reserved anycast address for any router on the network. Once delivered to a router, it replies using the most appropriate address.

In your case the Subnet Prefix is 0000:0000:0000:0000 (since RouterOS uses /64 prefix length by default) thus all addresses, other than 0000:0000:0000:0000::, are equally appropriate and RouterOS is at liberty to pick any one of them.

The 0000:0000:0000:0000:: is not discarded from address selection, but is less appropriate because in a network with multiple routers it does not uniquely identify the receiver.