I have three Virtual Machines: A, B, C.

Machine A and Machine C connected to each other through the B.

Their config:

A's IP: 2000::ffff:50a:b0b/120

C's IP: 2000::ffff:50a:d0d/120

Machine B has two interfaces - eth0 and eth1

IP of eth0 - 2000::ffff:50a:b0c/120

IP of eth1 - 2000::ffff:50a:d0e/120

I've set sysctl -w net.ipv6.conf.all.forwarding=1 at Machine B

Also, at machines A and C I've set default gateways:

At A

ip -6 ro add default via 2000::ffff:50a:b0c

At C

ip -6 ro add default via 2000::ffff:50a:d0e

I want to ping machine C from A and vice versa. The strange thing, that most part of ICMP packets are lost with message

Address unreachable

  • I instead use a bridge interface if the same subnet is on two different interfaces of a host – thrig May 9 '18 at 14:41
  • Which virtualization system are you using, and what type of interface (host, NAT, bridge, etc.) do the VMs have to the host? – ErikF May 9 '18 at 16:16
  • If the addresses in your examples are the actual addresses you are using, I would strongly suggest using Unique Local Addresses instead, and with /64 subnets instead of /120. – Johan Myréen May 9 '18 at 20:10
  • /120 means they are in different subnets, so it's routed. Reproducing the very same configuration with 3 network namespaces (linux) works fine here. – A.B May 11 '18 at 20:29
  • @A.B i don't know why, but the problem was in neighbor table. I manually set mapping between ip and mac addresses and it works fine – Evgeniy May 16 '18 at 10:10

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.