Opc us client - server Resources Bottleneck

Dears;

Hope to find you well.
I am experiencing some issues with a X20CP1585 CPU, wich with i am collecting some data through fieldbus and OPC UA any (40 tags) and then publish as an OPC UA server. Publish intervals are 500 msec.
Our opc ua clients was not able to reach the plc, even if ports are still active (telnet ok on 4840), therefore we are monitoring the number of sessions opened.
After restart the plc it went ok and was able to access SDM with following error:

I do not know if may be related, but i cannot see cpu utilization either:

The program is really minimal, just a copy paste from opc mapping tags (client) to publish and few alarms:

I may kindly ask your support and guidance to properly trouble shoot such issue starting maybe from such points.

Thanks in advance and BR;

Paolo

Hi Paolo,

Can you please send a screenshot of the System > Memory page in the SDM? This will show the memory status on the system.

When you say you restarted the PLC and it went ok, does this mean the OPCUA issues are now solved? Or what is the current issue? The error in the logger is specifically related to the SDM, not OPCUA. Any clarification to the current status would be helpful.

Thanks,
Rafael

1 Like

Dear Rafael;

Indeed. Here below:

How did you relate the error in the logger to SDM, is there a specific manual for logs? I did indeed search through help in AS (using 4.12 right now) but did not find andy mention.

The first issue may be related to the cpu not showing cpu utilization, so i may not be able to track any potential issue due to plc load:

image

It seems the OPC UA issues has been resolved after restarting the plc, or at least the issue not coming back that often.
As i am using Opc_Ua_Any to access some tags in a Siemens PLC OpcUaServer, i found that some of these tags was simply pointing at a wrong path (customer did change the path without letting us know).
At the moment i did reduce the opcUa tags accessed to the bare minimum, while trying to find a better use of diagnostic, and that is my second (i would not say issue, but suggestion).

Let us supposed the plc was not accessible and in a not defined state due to wrong adresses, number of subscription etc… How may i know if the number of tags implemented in opc Ua (both client and server side) is impacting in eccessive way on cpu resources? As far as i understood from the manual, cpu utilization is likely a simplified view of cycle tasks and idle times,Would cpu usage take in account this factor too?

Thanks for your support,

Best Regards;

Dears;

Today again we did lost the connection with plc in the morning and in the afternoon (the plc was not reachable online, through eternet IF2, therefore no diagnostic, same port that we use for opc connection ).
The cpu has been restarted and after that, we was able to reconnect both to the cpu and access to adress log.

We have the same cpu at desk, tomorrow we will try to trigger multiple opc ua connection in little time (or little overlap) and see if we can replicate the behaviour.

I may doubt and suppost the requests of opc server are too close one to each other:


With another X20CP1585 btw i have notice a similar behaviour, while cpu reached 100 percent utilization, it was not reachable till we restart it, so i may suppose so.
BR;
Paolo