Strange IIOP delay

[Petri]

We are using Borland VisiNotify in a development project. The VisiNotify server is running on a rp 5470 with UX 11.11 and latest (Dec '04) patches.

A self-written test tool triggers the server to issue a CORBA notification. Strangely, there appears to be a constant delay of 600 seconds (= 10 minutes ) before any IIOP traffic can be detected. The same VisiNotify server on a Linux host in the same subnet is running without any delay at all.

It seems the problem might be in the RP5470 configuration, as there is not much network equipment between, and the Linux host is working properly.
Any suggestion, what to look at?

Best regards,
/petri

[B. Hulst]

Hi,

This is corba, so having two separate systems using corba in the same subnet could mean trouble.

Both system can see and possibly use each's other objects if they have the same names.

It would be good to verify if both corba nameservices are running in each system on a different port number. Corba uses a network broadcast to locate the nameservice...and it would find two...the linux and the HPUX.

Use tcpdump (or similar) to sniff on that particular port the nameservice is using.
You might see interaction between the two systems. Unless you want that of course in a clusted environment. :-)

Regards,
Bob

[Petri]

Hi Bob, we have a federated CORBA Naming Service (there are more than these 2 machines!) and we do use tcpdump. Thanks for your reply, anyway.

[B. Hulst]

Hi,

There should be a borland tool called: osfind

/opt/Borlandxyzproduct/bin/osfind

Run this command and you will see all other object on a particular port on all hosts in the subnet.

I assume you have osfind. :-)

Regards,
Bob