Thank you for replying,

I will send you my code and the wireshark captures....But I will take
little time for doing that...

To give some more information about my application....please see my
comments below inlined....


On Nov 14, 2007 7:52 PM, Horvath, Elmer <[EMAIL PROTECTED]> wrote:
> Hi Jhankaar,
>
> A few questions:
> 1) Are you sending SOCK_RDM or SOCK_DGRAM?  The SOCK_DGRAM may get
> dropped at any point for any reason (congestion, queue limits, out of
> buffers)

I am sending on SOCK_RMD using sendmsg(). I checked with sendto it is
working fine.

>
> 2) Are you blasting things as fast as possible?  If you put in a slight
> delay on the send does this help?

There is a sleep of 1 second after every send.

>
> 3) If you only send multicast (and not the send to the non-existent
> address) does this fix things?

Yes, it works fine.
What I am seeing is....
When I do a unicast send to an invalid address then this address will
get set in the destnode field(ie 8th word) of the tipc header, which
is correct behaviour.
Now I do a send to the multicast address. But the destnode field
doesnt change to the new value because of that it doesnt reach the
receiver.


>
> It is strange that you are seeing your multi-cast not set the destnode
> field, though once you are fragmenting the packet, I believe you are
> seeing the header of the fragment and not the embedded multicast header.
When I saw the wireshark captured packet in case of fragmentaion, the
18th word(it is in options field of the header) was also same as
destnode. I am saying these specific field because when I send only
multicast packets, which get fragmented, I see these fields are zeros.

>
> It may be advisable to send different data to the unicast address and
> the multicast address to distinguish when you are capturing with
> wireshark.  I guess this will be self-evident when you post your code
> and wireshark data.

Yes I am sending different data to distinguish which packets I am
receiving and which I am not.... If my first packet is unicast then I
dont receive any packets and it is multicast packet then I am
receiving the first packet and then no packets I get.

This all is happening when I reach 1457 bytes of packet size. For 1456
I receive all multicast packets....

>
> Elmer
>
>
>
> -----Original Message-----
> From: [EMAIL PROTECTED]
> [mailto:[EMAIL PROTECTED] On Behalf Of
> jhankaar
> Sent: Wednesday, November 14, 2007 4:26 PM
> To: [email protected]
> Subject: [tipc-discussion] Seeing an issue with unicast and multicast
> onusing the same socket.
>
> Hi All,
>
> My Setup :
> I am using tipc 1.5.12.
> I am having 2 Redhat EL-4 machines. And on one machine I am running a
> sender and on the other one I am running a receiver.
> The receiver side is just receiving the packets and dropping it.
> The sender code is bit different. It is sending packets alternatively to
> 2 different addresses in a loop. One address is unicast address(TIPC
> name address) and the other one is a multicast address(TIPC nameseq
> address). The packet being sent to unicast address is a non-existant (
> or you can specify an address other than the above mentioned receiver's
> address). The above mentioned receiver falls in the multicast address
> group.
>
> The problem :
> Now, when I run my applications I see that the packets are being sent
> and received properly for 1456 or less bytes of data. But once I start
> sending packets of size bigger than 1456 the problem starts(when
> fragmentation for ethernet kicks in). I am not seeing the packets
> reaching the receiver...I saw through the Wireshark that the packets are
> reaching the receiver's machine but the TIPC is dropping the packets.
>
>
> I also saw that the tipc_multicast()  is not changing the destination
> address(8th word of TIPC header), set for previous send. And I tried
> putting zeros there explicitly  and the packets are reaching the
> receiver.
>
> Please correct me if I have assumed/doing something wrong.
> And also give me some hints on how should I go further.
>
> Thank You,
> Jhankaar
>
> ------------------------------------------------------------------------
> -
> This SF.net email is sponsored by: Splunk Inc.
> Still grepping through log files to find problems?  Stop.
> Now Search log events and configuration files using AJAX and a browser.
> Download your FREE copy of Splunk now >> http://get.splunk.com/
> _______________________________________________
> tipc-discussion mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/tipc-discussion
>

-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2005.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
_______________________________________________
tipc-discussion mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/tipc-discussion

Reply via email to