Just for Information - I cannot share the patch or anything.

My Corporate IT group just had us all roll back to 2.51 and some pacth to it from RA rather than 2.54 RSlinx.

They are stating "the OPC Server has an issue with 'locking up'".

Anybody else having OPC issue with RSlinx 2.54 may want to contact RA.

Can you share which OPC Client was in use, whether it was local or remote OPC, which type of controllers, what networks, what operating systems, and what symptoms the IT group calls "locking up" ?

I want to take a lot of IT guys and forcefeed them freshman philosophy books so they understand the basics of epistemology and stop treating "issues" like myth.

Yes. OPC will "lock up" due to a bug in DCOM (Microsoft bug) which happens if RS-Linx loses communication with a PLC for an extended period of time (long enough to fill up all the buffers in DCOM). Once the buffer overflow happens, it "locks up". I'm not sure that rolling back patches and/or OPC will even help. Rockwell is aware of the problem and you should be talking to them.

The solutions, if you want to call it that, is either to avoid RS-Linx as your OPC server altogether, use non-OPC protocols which don't have this problem (but DDE obviously isn't a real solution either), or else go to a competitor's OPC server.

Or, you have to CLOSE the connection to RS-Linx and then reopen it. I know, huge pain in the you-know-what. It is possible to use the status (@tags) through a separate OPC session to monitor things as long as whatever you do, do NOT access any of your real tags. Once you start getting real data, you're in trouble.

I've successfully managed to get Cimplicity PE to do the close/reopen trick. I think I've heard of others doing it with other HMI's but it's been so long since I worked with the others out there (and my experience with ControlLogix postdating the other HMI's) that I don't know how to fix the others.

If you're talking about the DCOM 108 bug, aka "RPC_E_Disconnected", then it affects all OPC products, not just RSLinx. And it doesn't affect Windows XP. And it would exist in any version of RSLinx.

So that might be the problem. It might not.

Bob, any more information you can post to narrow down this rumor ?

I'm sure I wouldn't be happy if I was a Rockwell support engineer who got the call saying "some guy on an Internet forum says there's a bug in RSLinx 2.54 and I have to roll back to 2.51 with some kind of patch. What do you have to say about that ?"

All I can add right now is that we were using Linx 2.54 to gather data from SLC 5/04's for InSQL and Intouch Wonderware Versions 9.5 and 10. We use Windows XP Pro SP2 and Server 2003 on our Data Collecton PC's and Servers.

We also have Logix 5000 and Logix 5 family processors, but use the Wonderware Native I/O Servers for them.