Hello Carsten,

I'm having problems to apply your patch series to my git repository. So I'm 
currently not able to test your patch with 2.0. :-(
But I'm of course willing to test it with 1.8. Could you please send all 
changed network files, so I can try to test that with 1.8, win7 and our current 
use case? Are there any differences between 1.8 and 2.0 for this that I need to 
consider?

Thanks,
Michael



Hello Michael,

On 11/01/2011 02:11 PM, Michael Raab wrote:
> I spent some time today looking at the ack handling between
> GroupMcastConnection and PointMcastConnection.
> Actually I don't think that it's much clearer to me now. :-(
> I would be very pleased if you may find some time to have deeper look.

I've done some work on this, but it's too early to commit. Also, I've 
used the OpenSG 2 sources, but there are very few (none?) differences in 
Source/Base/Network between the two, so once this code works reliable, 
backporting it to 1.x should be easy.
I'm attaching the patch series (quite messy, lots of work in progress 
commits) and the complete patch including the two test programs I've 
used (see below for instructions on how to run them).
I haven't tested any of this to actually run a cluster, only used the 
test programs have moved data back and forth.
Up to the last few commits in the series I could switch between old and 
new code and as a nice side effect I got higher throughput with the code :)

running servers (modes 0,1):
testEchoServer srvName
-g : mcast group for server discovery
-p : mcast port for server discovery
-t : test mode (use ONLY when running client in mode 2!)

running client (modes 0,1)
testEchoClient -t 0 -n 20 srvName0 srvName1 srvName2
-g : mcast group for server discovery
-p : mcast port for server discovery
-n : number of send/receive loops (0 for unlimited)
-t : test mode
-m : use mcast socket for data
-s : use stream socket for data

mode 0 : send variable sized block of random data and wait for
          servers to send same data back, compare
mode 1 : send variable sized block of random data, plus min,max,
          sum, let servers check min,max,sum
mode 2 : like mode 1, but randomly disconnect, reconnect to
          servers

This could use a lot more testing and I'll have to see if/how it works 
on windows. If you can give it a try and let me know how it fares that 
would be great.

        Cheers,
                Carsten





-- 
NEU: FreePhone - 0ct/min Handyspartarif mit Geld-zurück-Garantie!               
Jetzt informieren: http://www.gmx.net/de/go/freephone

------------------------------------------------------------------------------
RSA(R) Conference 2012
Mar 27 - Feb 2
Save $400 by Jan. 27
Register now!
http://p.sf.net/sfu/rsa-sfdev2dev2
_______________________________________________
Opensg-users mailing list
Opensg-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/opensg-users

Reply via email to