Zeros written to SQL while FactorySQL PC is being used

I have FactorySQL 4.2.10.0 running on an XPPro box along with other applications. The applications include RSLogix and RSViewME. The SQL database being written to is on a separate box. The problem seems to be while I am logged onto the FactorySQL box running other RS apps, FactorySQL seems to occasionally write zeros to the SQL Server for each Tag within a current record. So when I look at a historical trend such as with FactoryPMI, I see a trend that goes along just fine, but has occasioanl dips to zero for every tag. When I am not logged into the box or not running RS**, it appears to run just fine and I don’t see any of the zero dips. I either don’t see any error messages related to FactorySQL or I don’t know where to look. One symptom I do notice, is the PLC5 where all the apps are communicating to seems to be overloaded with ethernet traffic and I see very slow responses from some of the other HMI Panelview Plus’s and from RSView or RSLogix while I’m running them on the FactorySQL box.

Thanks
Rollie

If you’re using RSLinx as the OPC server, you can look under “DDE/OPC”->“Communication Events” to see if you’re getting any timeouts/errors during those times.

In FactorySQL, errors are visible under “Help->LogViewer”, though I doubt you’re getting any errors in this case. Instead, it’s likely that the group is simply getting new values with bad quality. However, getting 0s is a bit of a mystery, usually the groups will just get updated quality, and thus just hold on to the last values during that time. I suppose that the opc server could be sending 0 values along with the new qualities, but this would be fairly strange. The only other explanation I can come up with right now would be that you’re using “polled reads”, but I doubt that since it’s pretty rare.

Let us know if you find anything in the logs…

Regards,

The only timeout/errors found in the DDE/OPC Com Events were during known times when different plc’s were down, not during the times in question.

Factory SQL Logview did not have any errors during the questioned time.

I am not using any “polled reads” that I know of.

At this point I’m thinking of installing FactorySQL on a separate box then the one I run the RS apps on.

Do you actually see records with 0s in the database, or are you just going off of the trend dropping to 0 for a period of time? I ask because it’s possible the graph would show a break in data that might look like 0s if there just weren’t any records being inserted during that time.

If you go an look at the raw database data and see rows with 0s during that time, it may indicate that the OPC server is actually sending that to FactorySQL while under high load, which would be bad. If there are no records, that would also be bad, but probably not as much.

In general we’ve found that Kepware runs much more efficiently than RSLinx, so you might give that a try before buying another machine/copy of rslinx.

Regards,

Using a Query, I looked at the raw table. There really are Zeros in all the data fields except the t_stamp and *_ndx fields.

As for Kepware, I see it came with FactorySQL, so can I go ahead and install it and then if works, an activation license can be purchased?

Hi-

Yes, you can install it and set it up- it will run for 2 hours at a time, and you can restart it. And then if things work out, you can just call us up and buy it.

Regards,