Skip to content

System watchdog fw.v.: 2022.9.0

lukasHvlukasHv
edited May 2023 in PLCnext Technology & PLCnext Controls

Hello,

I have a problem with f2152 with firmware 20222.9.0 when I am in the debug mode and try trace a four values online in the same time call opcua method that is turning on the process I want to watch the plc goes down and reboot itself with warning message " system watchdog state changed: new state: UnacknowledgedWatchdog plc startup is: forbidden ".

Then I have to reboot plc one more time.

Have you any sugestions how to avoid this issue?

image.png


Comments

  • Eduard PLCnext TeamEduard PLCnext Team

    Hello Lukas,


    The appearance of HW-WD in debug mode indicates an error in the process. Could you please send me the answers to the following questions, please:

    1. You call the opcua method. Is it your self-implemented OPC UA client that contains the methods?
    2. If you don't start the OPC UA client, does the error occur?


    BR Eduard

  • lukasHvlukasHv

    Hello Eduard, 

    I am not sure about the 1st answer. The method is on the OPC UA server created according to guidance created by your colleague Martin. https://github.com/PLCnext/OpcUaMethods

    But the client is created by us with the use of Python library asyncua and it is calling the method with the help of the node string ID.

    image.png


     I checked the logs and it called the method only once at the time and there were no issues with connection prior to the accident.

    Second question: No and Yes I am 100% positive that the same issue will occure when the method is called from UaExpert. Becaues the CPU load is always high when I call any method.

    image.png

    The most of the program's blocks are on sescond core bcs it is less load during normal operations.

    image.png
    image.png

    The screenshots are without the debug mode.

  • Eduard PLCnext TeamEduard PLCnext Team
    edited May 2023

    Hello Lukas,

    the CPU-load of 85% on Core2 is high but not critical, the OPC-UA server operates in non-real time context and can't block the real-time program's. I also see the defined Task-WD's in your task configuration, if the program duration time exceeds the defined Task-WD-Time, the System will be stopped by Task-WD-Trigger and not by HW-WD/System-WD.


    "Second question: No and Yes I am 100% positive that the same issue will occure when the method is called from UaExpert. Becaues the CPU load is always high when I call any method."

    -> Please try to reproduce this issue with UaExpert (it's our reference tool). If the error is reproducible, we can analyse and fix it. In this case, please contact Phoenix Contact support department via pohone or e-mail address in your country.


    If your company location is in Germany: Kontakt | PHOENIX CONTACT

    Automatisierung und Safety

    Montag - Freitag: 08:00 - 16:00 Uhr

    +49 5235 3-11000

    support@phoenixcontact.de

    Thanks & BR Eduard

  • lukasHvlukasHv

    Hello,

    thank you for your time.

    I will try replicate this issue on our testbed next week.

  • lukasHvlukasHv

    We found the issue. 

    For some reason, our client opened a new connection for method calls instead of using an already-established connection for reading metrics. A connections for methods were properly closed but somehow it was capable of sending plc to crash as the screenshot in the first post here.

    Thank you Eduard for your time. 

  • Eduard PLCnext TeamEduard PLCnext Team

    Hi Lukas,

    no Problem & Thanks for the Feedback regarding the issue!

    BR Eduard

This discussion has been closed.