Factorytalk problems

Hi all.

We are starting up project with around 30000 points being
sampled by Cimplicity from RS Factory Talk through RS-OPC gateway.
On occasition the data appears to freeze. Cimplicity stops
receiving the unsolicited messages from the Rockwell products. The only way to release the freeze condition is to restart the server.
Cimplicity, RS Factory Talk, and RS-OPC all reside on the same
server.

Does anyone have input?

Thanks,

mbtoy

That sucks. Did this run successfully with fewer data points? You'd be well off posting what versions of the software and operating system you're using. What is the CPU/memory of the computer when this happens, specifically of the running services?

Are you sure Cimplicity speaks "Factory Talk"? I thought it would use RSLinx "Classic" (perhaps gateway version) through OPC. Can anyone please verify?

mbtoy said:

Hi all.

We are starting up project with around 30000 points being
sampled by Cimplicity from RS Factory Talk through RS-OPC gateway.
On occasition the data appears to freeze. Cimplicity stops
receiving the unsolicited messages from the Rockwell products. The only way to release the freeze condition is to restart the server.
Cimplicity, RS Factory Talk, and RS-OPC all reside on the same
server.

Does anyone have input?

Thanks,

mbtoy

Click to expand...

If your using Factory Talk Live data you could well be having DCOM errors.

There are a host of problems with this setup, even when using RSView SE which is designed to use Factory Talk.
There should be a diagnostics program installed on the server when you installed Factory Talk. Its called Diagnostics Viewer and is located under Rockwell Software.
Open that up and see what errors you are getting.

CPU and memory shouldnt matter as long as its current hardware and not some old donkey server but OS and the version of software you are running may make a big difference.

There are also a ton of patches on the Rockwell website. I would go and make sure everything is patched up to date.

If you are using RS Linx Classic, you may be suffering from "stale data" issue identified by RA 2-3 years ago. The OPC component stops updating data and freezes the old data. I believe there was a patch for this. Not certain if the patches we received were ever incorporated into the later versions of RS Linx classic. We have since moved to Linx enterprise so am uncertain if the problem still exists in classic.

I think I understand how your setup is configured, but you really need to post more information if you expect anything but pure speculation. Your best chance to resolve this problem is through direct contact with Rockwell Automation and GE.

For those speculating along: RSLinx Enterprise is the data engine that supplies connectivity between the FactoryTalk LiveData service and Rockwell Automation controllers. We can safely assume that's what mbtoy is running.

FactoryTalk OPC Gateway is a relatively new (2 years) product that allows ordinary OPC clients (instead of FactoryTalk LiveData clients) to use RSLinx Enterprise.

Sheldn is probably referring to the notorious "DCOM 108" problem, which RA and Microsoft rooted out in Windows 2000 and 2003 Server operating systems about the same time Microsoft announced they were going to drop DCOM support from Windows Vista anyhow. I cannot rule this out. The repairs for "DCOM 108" are Windows patches, and the problem is well described in RA Knowledgebase Answer ID # 37039.

I suggest two diagnostic approaches.

The first is to use the FactoryTalk Diagnostic Log, which SLC_Integrator recommended.

The second is to run the OPC Test Client, which is included with RSLinx Classic and probably also bundled with FactoryTalk OPC Gateway. This will tell you if all OPC connectivity has failed, or if it's just the connections between FactoryTalk OPC Gateway and your Cimplicity client.

My third suggestion is that you may not find the problem in the RSI software or in DCOM. In my experience with very large Cimplicity installations (at a certain microprocessor industry leader whose name rhymes with "Schmintel"), Cimplicity clients would disconnect themselves from running OPC Servers (both RA and others) and require reboot. That was about four years ago, but I don't think it's beyond possibility.