вт, 31 июл. 2018 г. в 4:35, Jason Gunthorpe <jgg@xxxxxxxx>:
>
> On Tue, Jul 31, 2018 at 01:57:42AM +0300, Vasiliy Tolstov wrote:
> > пт, 27 июл. 2018 г. в 18:38, Jason Gunthorpe <jgg@xxxxxxxx>:
> > >
> > > On Fri, Jul 27, 2018 at 12:42:47PM +0300, Vasiliy Tolstov wrote:
> > > > пт, 20 июл. 2018 г. в 11:31, Vasiliy Tolstov <v.tolstov@xxxxxxxxx>:
> > > > >
> > > > > > I think that who first get ipv6 address wins and second have error.
> > > > > > I'm disable DAD and now server 1 and server 2 have identical ipv6
> > > > > > slaac addresses. And i know that this is not bird problem and not
> > > > > > dnsmasq problem because in case of plain ethernet devices with normal
> > > > > > hwaddres this issue not happened.
> > > > > >
> > > > >
> > > > > Gentle Ping....
> > > > >
> > > >
> > > > Any chance to get this fixed or this is 100% dnsmasq/bird ... problem?
> > > > (that does not appeared when use ethernet device)
> > >
> > > Unless you can show the kernel itself is producing the address wrong
> > > we can't really help you.. The kernel code looks fine to me, unless
> > > something broke recently.
> > >
> > > Test using simply radvd and reboot everything.
> > >
> > > Jason
> >
> > radvdump recieved on server1 and server2:
> > interface bond0
> > {
> > AdvSendAdvert on;
> > # Note: {Min,Max}RtrAdvInterval cannot be obtained with radvdump
> > AdvManagedFlag off;
> > AdvOtherConfigFlag off;
> > AdvReachableTime 0;
> > AdvRetransTimer 0;
> > AdvCurHopLimit 64;
> > AdvDefaultLifetime 300;
> > AdvHomeAgentFlag off;
> > AdvDefaultPreference medium;
> >
> > prefix fd00:1::/64
> > {
> > AdvValidLifetime 86400;
> > AdvPreferredLifetime 14400;
> > AdvOnLink on;
> > AdvAutonomous on;
> > AdvRouterAddr on;
> > }; # End of prefix definition
> >
> > }; # End of interface definition
> >
> > i'm test with radv:
> > server1 dmesg:
> > [1420250.782673] ICMPv6: RA: ndisc_router_discovery failed to add default route
> > [1420251.074430] IPv6: bond0: IPv6 duplicate address
> > fd00:1::8200:2ff:fe09:fe80 used by 00:00:00:00:00:00 detected!
> >
> > server2 dmesg:
> > clean
> >
> > but server2 have two ipv6 addresses:
> > 6: bond0: <BROADCAST,MULTICAST,MASTER,UP,LOWER_UP> mtu 2044 state UP qlen 1000
> > inet6 fd00:1::8200:2ff:fe09:fe80/64 scope global dynamic
> > mngtmpaddr noprefixroute
> > valid_lft 86393sec preferred_lft 14393sec
>
> No idea what this is. Did you reboot everything and stop other
> advertising servers? I think you need to tcpdump to investigate this..
>
> > inet6 fd00:1::221:2800:1a0:cfe2/64 scope global dynamic mngtmpaddr
> > valid_lft 86393sec preferred_lft 14393sec
> > inet6 fe80::221:2800:1a0:cfe2/64 scope link
> > valid_lft forever preferred_lft forever
>
> This looks like it worked to me, the above is the radv address and
> the link local address, both with the correct suffix.
>
> Jason
Thanks for answer, i'm double check all things. Invalid ipv6 address
generated by kernel on server2, tcpdump:
tcpdump -vvvv -ttt -i bond0 icmp6 and 'ip6[40] = 134'
tcpdump: listening on bond0, link-type LINUX_SLL (Linux cooked),
capture size 262144 bytes
00:00:00.000000 IP6 (flowlabel 0xa0d39, hlim 255, next-header ICMPv6
(58) payload length: 48) _gateway > ff02::1: [icmp6 sum ok] ICMP6,
router advertisement, length 48
hop limit 64, Flags [none], pref medium, router lifetime 300s,
reachable time 0ms, retrans time 0ms
prefix info option (3), length 32 (4): fd00:1::/64, Flags
[onlink, auto, router], valid time 86400s, pref. time 14400s
0x0000: 40e0 0001 5180 0000 3840 0000 0000 fd00
0x0010: 0001 0000 0000 0000 0000 0000 0000
And address appeared only then i'm run radvd on gateway node.
Also i'm disable accept_dad = 0 on server1 and it receive invalid ipv6
too (fd00:1::8200:2ff:fe09:fe80/64)
Also i'm double check - not RA sended and received by this servers. --
Vasiliy Tolstov,
e-mail: v.tolstov@xxxxxxxxx
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
