RE: Connection loss with Rockwell's RSSQL interface

Hi Michael
Try turning in sqlnet logging on the system where the RSSQL server is
running, then you can see all your traffic in and out of that server and
should be able to see the responses coming back.
Did this happen only after upgrading the database from 8 to 10 or did
some other event cause this to happen? There are a lot of differences in
10g to 8i. 
What version is the oracle client on the RSSQL server? 
Does RSSQL have recommendations for oracle client versions? 


Cheers

 

-----Original Message-----
From: Michael Ray [mailto:topshot.rhit@xxxxxxxxx] 
Sent: Tuesday, 27 March 2007 08:02 AM
To: oracle-l@xxxxxxxxxxxxx
Subject: Connection loss with Rockwell's RSSQL interface

We've been having some intermittent connection loss/hang/timeout issues
with PLCs on the production floor that cause significant downtime. PLCs
communicate via RS Linx from what I'm told to a Rockwell Software RSSQL
server (OPC server which ties a "tag" in the PLC to a field, procedure,
etc in the database), which then translates that request and does the
communication with Oracle via either OCI or ODBC. The database was 8i
and is now 10g.

It seems the only thing we know for certain is the PLC isn't getting the
"done" signal back from the RSSQL server so it just sits and waits. No
explicit error messages appear anywhere. The RSSQL server is capable of
knowing when a PL/SQL proc failed somehow, but that's not happening.

How can I prove without a doubt that the response from the Oracle server
makes it to the RSSQL server? Both Oracle and RSSQL are running on
Windows machines and neither are resource starved.

Perhaps a network (packet) sniffer is the only way to solve it? Are
there any good free software-based ones? We don't have the budget for a
hardware sniffer.

Peace,
Michael

--
http://www.freelists.org/webpage/oracle-l


--
http://www.freelists.org/webpage/oracle-l

Other related posts: