I am still waiting for your comments. Can you put some comment to my past
update?
Thanks.


2008/10/23 Bogon Choi <[EMAIL PROTECTED]>

> I tested with Tcpdump. I have two interfaces, eth0 and eth2 (gigabit).
> When I test with eth0 and its front-end switch, perfectly UDP multicast
> works. By the way, when I test with eth2 and gigabit back-end switch, it
> still does not work.
>
> So, I ran tcpdump to see packets on eth2 of both sides. The interesting
> thing is that I can see the multicast packets of both sides, eth2s. From
> this situation, I don't know why two controllers cannot receive those
> packets through the interface.
>
> So, I assume that the controller does not know how to listen from the
> second interface, eth2. Maybe, it can only listen from the first one, eth0,
> for UDP communication. Hard coded something.
>
> Totally I don't understand the current situation. Packets are floating
> around on the gigabit network. But, the packets can not be received on each
> controller. Can you give me some idea? Also, I want to mention that the
> JGroups simple test worked with the gigabit switch.
>
> In the worst case, I have in mind switching the order of interfaces, making
> gigabit interface to eth0.
> Thanks.
>
>
> 2008/10/22 Bogon Choi <[EMAIL PROTECTED]>
>
> Thanks for your notice.
>> I tested with JGroups sample. But, it worked with the current routing
>> table.
>> And, today I sent an email to the network administrator to check out the
>> physical switch setting.
>> As you said, it might filter UDP milticasting packets. Thanks anyway.
>>
>> I am waiting from a reply from the admin. And, in the worst case, I have
>> to setup tcpdump to trace all network stacks and routes.
>>
>>
>>
>> On Wed, Oct 22, 2008 at 10:55 AM, Emmanuel Cecchet <[EMAIL PROTECTED]>wrote:
>>
>>> Hi,
>>>
>>>> After showing this error, two controllers are partitioned.
>>>>
>>> Actually they never saw each other and they were never in a group.
>>> Make sure that your switch supports UDP multicast. You can use the
>>> JGroups demo and test your network config until the demo works. Once you
>>> have your network settings right, you can try to use Sequoia.
>>> If you are using Linux, make sure that your kernel supports IP multicast.
>>> Use a tool like tcpdump or ethereal to check how packets are routed on
>>> your machine.
>>>
>>> Hope this helps,
>>> Emmanuel
>>>
>>>
>>>> I am using Appia for UDP multicasting. When I type "enable <backend>"
>>>> from the other controller,
>>>> the prompt is stuck and the console shows a bunch of
>>>> nakfifo.multicast.NakFifoMulticastSession nacked.
>>>> And then, controllers are partitioned each other. They don't see each
>>>> other finally.
>>>>
>>>> I was testing how to replicate a virtual database to two controllers.
>>>> Each virtual database conf. was setup to use one backend of a different
>>>> node, with RAIDb-1 type.
>>>>
>>>> First time, I tries to solve this problem by using 'JGroups'. But,
>>>> jgroups also was stuck at 'enabling the backend'.
>>>> So, I moved to 'Appia'. Appia seemed to work with TCP setup. Yes, it
>>>> worked. But, it does not work with UDP multicasting.
>>>>
>>>> I turned off firewalls. So, this is not firewall problem.
>>>>
>>>> I added the entry in routing table, such as:
>>>> /sbin/route add -net 224.0.0.0 <http://224.0.0.0> netmask 240.0.0.0 <
>>>> http://240.0.0.0> dev eth2
>>>>
>>>> Following is my routing table of a controller
>>>>
>>>> Kernel IP routing table
>>>> Destination     Gateway         Genmask         Flags Metric Ref    Use
>>>> Iface
>>>> 10.60.10.0 <http://10.60.10.0>      *               255.255.255.0 <
>>>> http://255.255.255.0>   U     0      0        0 eth0
>>>> 192.168.10.0 <http://192.168.10.0>    *               255.255.255.0 <
>>>> http://255.255.255.0>   U     0      0        0 eth2
>>>> 224.0.0.0 <http://224.0.0.0>       *               240.0.0.0 <
>>>> http://240.0.0.0>       U     0      0        0 eth2
>>>> default         10.60.10.1 <http://10.60.10.1>      0.0.0.0 <
>>>> http://0.0.0.0>         UG    0      0        0 eth0
>>>>
>>>> Actually, I wanted to use gigabit ethernet for all products
>>>> communication. So, I forwarded gossip and udp packets to eth2, gigbit
>>>> ethernet. When I used TCP setup, it used eth0 wholly. What is the problem?
>>>>
>>>> ------------------------------------------------------------------------
>>>>
>>>> _______________________________________________
>>>> Sequoia mailing list
>>>> [email protected]
>>>> https://forge.continuent.org/mailman/listinfo/sequoia
>>>>
>>>
>>>
>>> --
>>> Emmanuel Cecchet
>>> FTO @ Frog Thinker Open Source Development & Consulting
>>> --
>>> Web: http://www.frogthinker.org
>>> email: [EMAIL PROTECTED]
>>> Skype: emmanuel_cecchet
>>>
>>> _______________________________________________
>>> Sequoia mailing list
>>> [email protected]
>>> https://forge.continuent.org/mailman/listinfo/sequoia
>>>
>>
>>
>>
>> --
>> 여호와는 네게 복을 주시고 너를 지키시기를 원하며
>> 여호와는 그 얼굴을 네게 비추사 은혜 베푸시기를 원하며
>> 여호와는 그 얼굴을 네게로 향하여 드사 평강 주시기를 원하노라
>> (민수기 6:24-26)
>>
>
>
>
> --
> 여호와는 네게 복을 주시고 너를 지키시기를 원하며
> 여호와는 그 얼굴을 네게 비추사 은혜 베푸시기를 원하며
> 여호와는 그 얼굴을 네게로 향하여 드사 평강 주시기를 원하노라
> (민수기 6:24-26)
>



-- 
여호와는 네게 복을 주시고 너를 지키시기를 원하며
여호와는 그 얼굴을 네게 비추사 은혜 베푸시기를 원하며
여호와는 그 얼굴을 네게로 향하여 드사 평강 주시기를 원하노라
(민수기 6:24-26)
_______________________________________________
Sequoia mailing list
[email protected]
https://forge.continuent.org/mailman/listinfo/sequoia

Reply via email to