Re: problems on eventchannel

Hi,

That old Event Service implementation was not able to cope with clients 
that died without first calling disconnect. It would continue trying to 
send events to the dead client forever. I think that this is the 
problem you are describing.

Try the newer version of omniEvents (2.6.1) at 
http://www.omnievents.org/ - it is much more robust. (Beware though - 
the format of the persistency database has changed - make sure you 
delete your omnievents-HOSTNAME.log/bak files before you upgrade.)

In this case, when the client dies, omniEvents will notice, and not 
bother trying to deliver any more events to it.

-Alex

--
:: alex tingle
:: http://www.firetree.net/consulting/
:: alex.tingle AT firetree.net  +44-7901-552763

On 18 May 2005, at 14:36, DIEZ Ignacio wrote:

> Hello,
> I'm having problems with my OmniORB Event Channel CORBA.
> I'm using omniORB 3.0.3 and omniEvents 2.1.2.
> I have 1 server that starts a Naming Service and an Event Channel and
> 3 or 4 clients that are connected to this server. The problem is that
> when a client dies, the server polls over all of his own ports trying
> to find the client.
> So, If the server connects through the 1234 port with a client in a
> 6789 port and dies, then the server will ask to the client from 1235
> to 6789, 1236 to 6789 and so on...
> The real problem is that when the client restarts.. the server uses
> another port to connect with the client, but the server continues 
> trying
> to connect with the older ports of the clients.
> The result is a lot of traffic (TCP/IP packets) over the network.
> The Event Channel is started with -ORBInitRef
> NameService=corbaloc::1.0@pcproof:1235/NameService
> -ORBpoa_iiop_name_port localhost:1450 -ORBinConScanPeriod 3
> -ORBoutConScanPeriod 3.
> Could anybody help me, please??
>
> _______________________________________________
> omniORB-dev mailing list
> [email protected]
> http://www.omniorb-support.com/mailman/listinfo/omniorb-dev