SQL Tags and Stale Data

Came into work this morning and the MySQL DB was faulted, or so my FPMI told me it was. I couldn’t get MySQL to shut down and restart, so I restarted the computer. Once restarted, all of my SQL tags showed “STALE”. I tried a couple of things, none of which worked well. I had a couple of instances where I could edit the scan class and THAT would work, but not for all of them. Then I exported all the tags, changed all of their scan classes. Deleted all the SQL tags in designer and then uploaded my changed csv. NOW all of them are stale again and NOTHING seems to get them working again. Please help. We have NO alarm log now LOL…

-Chris Sanders

Wow, where to start? … how about you give us a call?

Just as a followup, this problem (the STALE tags anyways) apparently resolved itself after FactorySQL finished running through its datacache.

Alright, I’m still getting stale tags once in awhile. I DEFINATELY notice it when i’m running a query on a LARGE table. But sometimes, my sql tagged objects flash stale for about 5 sec then go back GOOD. Any ideas? Perhaps my fsql groups are not set up EFFICIENTLY? Or perhaps the computer has too large of a workload with OTHER software running (ie: rslogix[4 instances], rslinx)?

Hey Chris,
First off, can we have some information regarding the query, and the size and layout of the table. That could be the cause of the issue right there, a long running script preventing FSQL from updating the table, caching the data, then hammering the data back in once the lock is relieved. Also, what database are you using. Also, server specs?

Another thing that catches my eye, is RSLinx. Are you using this as your OPC server. If so, you can alleviate a ton of CPU usage by switching to a server from Kepware of Matrikon. For example, we were running RSLinx/IA on a facility, and the CPU utilization from RSLinx alone was around %50. Kepware, used 3%. It has to do with the type of OPC spec they use, and how they get the data. In a nutshell, avoind RSLinx as an OPC server like the plague. Plus, the RSLinx version you need to use it as a server costs more money then a solution from other OPC vendors.

Its pretty bad that when you go to Rockwell courses, they mock how bad RSLinx is.

This machine is CURRENTLY a troubleshooting computer. At most times, it runs 4 instances of rslogix 5, connected to 4 PLC’s of course. And starting today, I noticed that the tags go stale when I open up FPMI designer. Then shortly, they will all refresh and become good.