[omniORB] Blockage on a heavy network

[Thierry DESCOMBES]

I have noticed a strange thing too: All is working fine if the PC and the 
etrax are plugged on a Cisco3524XL switch, but when I replace it by a cheap 
Dlink DES-1024D switch, the blocage occurs after less than 30 minutes 
(whether I use OmniORB 3.0.4 or 4.0.1) !!

Cheers,
DESCOMBES Thierry

At 13:35 09/06/2003 +0100, Duncan Grisby wrote:
>On Wednesday 4 June, Thierry DESCOMBES wrote:
>
> > I have ported OmniORB 4.0.1 to the Axis Etrax 100LX
> > (http://developper.axis.com) - I can give you the file Etrax.mk I made, if
> > you want to add it in a next release.
>
>Yes, please.
>
> > It works quite fine. I have 5 Etrax communicating all with a single PC. 
> The
> > PC loops and calls, each time, simultaneously on the 5 Etrax a function
> > which return 26Kbytes of datas. This works fine if the PC is on a Gigabit
> > port of the switch. But when I plug it on a 100Mbits ports, the program
> > starts and works fine during 30 minutes maximum and then blocks: The
> > function called on one of the Etrax (it's never the same) is blocked 
> before
> > returning the datas, and the PC wait infinitely.
> > I was using OmniORB 3.0.4 and I found out the same behaviour.
> > I don't understand at all what happens, since, according to my
> > calculations, the rate of datas doesn't exceed 7Mbits in average !
>
>Please get a trace from the server with -ORBtraceLevel 25
>-ORBtraceThreadId 1.  That will help us see where it is blocking.
>
>Cheers,
>
>Duncan.
>
>--
>  -- Duncan Grisby         --
>   -- duncan at grisby.org     --
>    -- http://www.grisby.org --