ab double team ping6 issue

[aloughlam]

Hi guys,

In active backup double team, there has been an issue with first ping6 being unreachable. this was the case since we started testing ipv6 but I was more focus on throughput and I missed this issue. I have tried -327.el7 and -380.el7 and different versions of lnst and the results are the same "unreachable". Also w/run or w/config_only.

w/ LNST
on server1
if I "ping6 fc00::2 -I fc00::2" it is unreachable (keep ping running)

on server2:
if I "ping6 fc00::1" or I do "tcpdump" from server2, ping6 on server1 starts passing

Aaron and I had spent sometime troubleshooting and we came to the conclusion that server2 is the problem and refuses to reply to server1 neighbour solicitation

so I took lnst out of the equation:
I have done a fresh rhel7 install and created team0 manually on both servers using:

• ip link set p5p1 down
• ip link set p5p2 down
• teamd -d -t team0 -c '{ "runner" : { "name": "activebackup" }, "ports" : { "p5p1": { "link_watch": {"name": "ethtool"}}, "p5p2": { "link_watch": {"name": "ethtool"}}}, "hwaddr": "00:11:22:33:44:55" }'
• ip -6 addr add fc00::1/64 dev team0
• ip addr add 192.168.0.1/24 dev team0
• ip link set team0 up

both ping and ping6 pass,