Andrew,
please make really sure that mcast works. You can use for example omping
for that.
What seems to be strange is especially node A. If you compare your node
A output:
> 10:33:57.364852 IP 10.40.129.209.5464 > 224.0.0.187.5465: UDP, length 82
> 10:33:57.599081 IP 10.40.129.209.5464 > 224.0.0.187.5465: UDP, length 82
with node B output:
> 10:33:57.612370 IP 10.40.129.210.5464 > 224.0.0.187.5465: UDP, length 112
> 10:33:57.612473 IP 10.40.129.209.5464 > 224.0.0.187.5465: UDP, length 156
you can see node B sees both 10.40.129.210.5464 and 10.40.129.209.5464,
but node A sees only 10.40.129.209.5464.
You can also try to turn on debugging (logging debug: on in config file)
and see logs.
Regards,
Honza
and k napsal(a):
> Hi everybody,
>
> I've got a problem with join two nodes with corosync when using multicast.
>
> I've got 2 nodes:
>
> A:10.40.129.209
> B:10.40.129.210
>
> interface in corosync.conf
>
> interface {
> # The following values need to be set based on your
> environment
> ringnumber: 0
> bindnetaddr: 10.40.128.0
> mcastaddr: 224.0.0.187
> mcastport: 5465
> }
>
>
> Here is tcpdump on node A.
>
> At first corosync is up on node A, I see it sends messages to multicast:
>
> 10:33:55.724608 IP 10.40.129.209.5464 > 224.0.0.187.5465: UDP, length 82
> 10:33:55.958984 IP 10.40.129.209.5464 > 224.0.0.187.5465: UDP, length 82
> 10:33:56.193231 IP 10.40.129.209.5464 > 224.0.0.187.5465: UDP, length 82
> 10:33:56.427601 IP 10.40.129.209.5464 > 224.0.0.187.5465: UDP, length 82
> 10:33:56.661914 IP 10.40.129.209.5464 > 224.0.0.187.5465: UDP, length 82
> 10:33:56.896153 IP 10.40.129.209.5464 > 224.0.0.187.5465: UDP, length 82
> 10:33:57.130481 IP 10.40.129.209.5464 > 224.0.0.187.5465: UDP, length 82
> 10:33:57.364852 IP 10.40.129.209.5464 > 224.0.0.187.5465: UDP, length 82
> 10:33:57.599081 IP 10.40.129.209.5464 > 224.0.0.187.5465: UDP, length 82
>
> Now I stared corosync on instance B
>
> 10:33:57.612370 IP 10.40.129.210.5464 > 224.0.0.187.5465: UDP, length 112
> 10:33:57.612473 IP 10.40.129.209.5464 > 224.0.0.187.5465: UDP, length 156
> 10:33:57.681781 IP 10.40.129.209.5464 > 224.0.0.187.5465: UDP, length 156
> 10:33:57.750903 IP 10.40.129.209.5464 > 224.0.0.187.5465: UDP, length 156
> 10:33:57.820029 IP 10.40.129.209.5464 > 224.0.0.187.5465: UDP, length 156
> 10:33:57.823920 IP 10.40.129.210.5464 > 224.0.0.187.5465: UDP, length 82
> 10:33:57.890109 IP 10.40.129.209.5464 > 224.0.0.187.5465: UDP, length 156
> 10:33:57.959233 IP 10.40.129.209.5464 > 224.0.0.187.5465: UDP, length 156
> 10:33:58.028354 IP 10.40.129.209.5464 > 224.0.0.187.5465: UDP, length 156
> 10:33:58.058154 IP 10.40.129.210.5464 > 224.0.0.187.5465: UDP, length 82
> 10:33:58.098350 IP 10.40.129.209.5464 > 224.0.0.187.5465: UDP, length 156
> 10:33:58.167477 IP 10.40.129.209.5464 > 224.0.0.187.5465: UDP, length 156
> 10:33:58.236598 IP 10.40.129.209.5464 > 224.0.0.187.5465: UDP, length 156
> 10:33:58.292490 IP 10.40.129.210.5464 > 224.0.0.187.5465: UDP, length 82
> 10:33:58.306685 IP 10.40.129.209.5464 > 224.0.0.187.5465: UDP, length 156
> 10:33:58.375919 IP 10.40.129.209.5464 > 224.0.0.187.5465: UDP, length 156
> 10:33:58.445041 IP 10.40.129.209.5464 > 224.0.0.187.5465: UDP, length 156
> 10:33:58.514163 IP 10.40.129.209.5464 > 224.0.0.187.5465: UDP, length 156
> 10:33:58.526730 IP 10.40.129.210.5464 > 224.0.0.187.5465: UDP, length 82
> 10:33:58.583924 IP 10.40.129.209.5464 > 224.0.0.187.5465: UDP, length 156
> 10:33:58.653048 IP 10.40.129.209.5464 > 224.0.0.187.5465: UDP, length 156
> 10:33:58.666071 IP 10.40.129.210.5464 > 224.0.0.187.5465: UDP, length 38
> 10:33:58.722251 IP 10.40.129.209.5464 > 224.0.0.187.5465: UDP, length 156
> 10:33:58.791379 IP 10.40.129.209.5464 > 224.0.0.187.5465: UDP, length 156
> 10:33:58.860500 IP 10.40.129.209.5464 > 224.0.0.187.5465: UDP, length 156
>
> As you can see mutlicast traffic is visible on both nodes, but cluster wont
> to join, it keep sending packet of length 82 and 156 continously.
>
>
> Any idea how can I debug this strange situation ? In logs there is no any
> errors, it seem that nodes doesnt see each other, but multicast traffic is
> visible on both.
>
> What is more, as you can see above, when i run second instance corosync on
> node B, node A see it because it send diffrent mesage of length 156 b.
>
> Any help ?
>
> Thank in advance
> Andrew
>
>
>
> _______________________________________________
> discuss mailing list
> discuss@xxxxxxxxxxxx
> http://lists.corosync.org/mailman/listinfo/discuss
>
_______________________________________________
discuss mailing list
discuss@xxxxxxxxxxxx
http://lists.corosync.org/mailman/listinfo/discuss
