We reproduce the issue on this diagram: client0 (netns) --> ls1 --> lr1 --> ls2 --> server0 (netns) "gateway_mtu =1300" is set on logical_router_port lr1ls2 (towards server side) On old version: [root@wsfd-advnetlab35 bz1846300]# rpm -qa | egrep "ovn|openv" openvswitch2.13-2.13.0-29.el7fdp.x86_64 openvswitch-selinux-extra-policy-1.0-15.el7fdp.noarch ovn2.13-2.13.0-34.el7fdp.x86_64 ovn2.13-host-2.13.0-34.el7fdp.x86_64 ovn2.13-central-2.13.0-34.el7fdp.x86_64 We noticed, - There was no MTU too big notification for ipv6. - The ICMP packet that server received is not fragmented and it is same as original full size. (packet#12 in reproduced_oldversion.pcap attached) 2000::1 2001::1 ICMPv6 1362 Echo (ping) request id=0xf0c1, seq=1, hop limit=63 (reply in 13) On new version: [root@wsfd-advnetlab35 bz1846300]# rpm -qa | egrep "ovn|openv" openvswitch2.13-2.13.0-29.el7fdp.x86_64 ovn2.13-central-20.06.1-2.el7fdn.x86_64 openvswitch-selinux-extra-policy-1.0-15.el7fdp.noarch ovn2.13-host-20.06.1-2.el7fdn.x86_64 ovn2.13-20.06.1-2.el7fdn.x86_64 We noticed, - There is a notification of mtu too big when we send the first ping From 2000::a icmp_seq=1 Packet too big: mtu=1300 - ICMP6 packets that server received after the first ping are fragmented. (packet#12 and 13 in repro_fixed_newversion.pcap attached) 2000::1 2001::1 IPv6 1310 IPv6 fragment (off=0 more=y ident=0xf262b866 nxt=58) 2000::1 2001::1 ICMPv6 122 Echo (ping) request id=0xf9c9, seq=1, hop limit=63 (reply in 14) As below is the first ping results on new version (20.06.1-2.el7fdn). + ip netns exec client0 ping 192.168.2.1 -c 1 PING 192.168.2.1 (192.168.2.1) 56(84) bytes of data. 64 bytes from 192.168.2.1: icmp_seq=1 ttl=63 time=1001 ms --- 192.168.2.1 ping statistics --- 1 packets transmitted, 1 received, 0% packet loss, time 0ms rtt min/avg/max/mdev = 1001.689/1001.689/1001.689/0.000 ms + ip netns exec client0 ping 192.168.2.1 -c 1 -s 1300 PING 192.168.2.1 (192.168.2.1) 1300(1328) bytes of data. From 192.168.1.254 icmp_seq=1 Frag needed and DF set (mtu = 1300) --- 192.168.2.1 ping statistics --- 1 packets transmitted, 0 received, +1 errors, 100% packet loss, time 0ms + ip netns exec client0 ping6 2001::1 -c 1 PING 2001::1(2001::1) 56 data bytes --- 2001::1 ping statistics --- 1 packets transmitted, 0 received, 100% packet loss, time 0ms + ip netns exec client0 ping6 2001::1 -c 1 -s 1300 PING 2001::1(2001::1) 1300 data bytes From 2000::a icmp_seq=1 Packet too big: mtu=1300 --- 2001::1 ping statistics --- 1 packets transmitted, 0 received, +1 errors, 100% packet loss, time 0ms The reproduce script reproduce_script.sh is also attached below.
Created attachment 1703121 [details] reproduce commands
Created attachment 1703122 [details] pcap - old ovn version - ipv6 is not fragmented
Created attachment 1703123 [details] pcap - on new ovn version - ipv6 is fragmented
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory (ovn2.13 bug fix and enhancement update), and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHBA-2020:3488
*** Bug 1839847 has been marked as a duplicate of this bug. ***