Hello,

you have a license purchased on 5/23/2022 without upgrade assurance. That means that it only covers the versions released max. 60 days after the purchase date. So, according to kb.opclabs.com/Versions , this comes to maximum version 5.63 (2022.1).

We only make bug fixes to the current version, that's why it has been made into QuickOPC 5.70 (2022.2).

Normally you would therefore need to purchase an upgrade in order to use QuickOPC 2022.2. But since you were instrumental in finding this "valuable" bug, you will get a kind of bounty, and you will receive a license key that covers version 2022.2. I will send it to you by email.

Best regards

Hello,
I was able to get the event flow now. And, I was able to reproduce the memory leak.

It should be now fixed. The fix is in QuickOPC 5.70.1012.1 or later, which is on our Web site, and on nuget.org.
Please rebuild with this new version, and retest.

Thank you for reporting the issue!

Best regards

Hello,
still no flow of events from the server.

Regards

Hello.

The OPC UA server on opc.tcp://opcua.adnet.hr:81 does not give me any events, except for a single condition that remains static (in initial Refresh). This is true both in my test client code with QuickOPC, AND using UaExpert, as shown on following pictures.






If there is a different notifier node to be used, or some other trick to make it work, let me know.

Regards

Hello.

I was unable to reproduce the problem. Here is what I did:

1. I took your console-based program.
2. Made a special change to our currently released version (2022.2) to generate 1000 identical event notifications (instead of just 1) at the point we received them from the OPC UA stack, so that I can simulate a higher rate of notifications than what the server actually provides.
3. Changed it to reference the above special build of our currently released version.
4. Change the endpoint URL to point to my local Prosys simulation server.
5. Commented out the unused data change handler, only kept the event notification handler.
6. Ran the program and monitored the Private Bytes using PerfMon.

Memory consumption chart of cca 12 hours run:



For record, the program now looks like this:

 

 

using OpcLabs.EasyOpc.UA;

using OpcLabs.EasyOpc.UA.AddressSpace.Standard;

using OpcLabs.EasyOpc.UA.AlarmsAndConditions;

using OpcLabs.EasyOpc.UA.OperationModel;

using System;

 

namespace OpcLabsBug

{

 internal class Program

 {

  private static EasyUAClient client;

 

  [MTAThread]

  static void Main( string[] args )

  {

   client = new EasyUAClient();

 

   subscribe();

 

      Console.ReadLine();

    }

 

  private static void subscribe()

  {

   UAEndpointDescriptor endpointDescriptor = new UAEndpointDescriptor( "opc.tcp://localhost:53530/OPCUA/SimulationServer" );

   if ( client == null )

    client = new EasyUAClient();

 

   client.EventNotification += Client_EventNotification;

   UAAttributeFieldCollection uAAttributeFields = UABaseEventObject.AllFields;

   int eventMonitoredItem = client.SubscribeEvent(

    endpointDescriptor,

    UAObjectIds.Server,

    0,

    uAAttributeFields );

  }

 

  //private static void ProcessDataChange( object sender, EasyUADataChangeNotificationEventArgs eventArgs )

  //{

  // if ( eventArgs.Succeeded )

  // {

  //  string recordId = eventArgs.Arguments.NodeDescriptor.NodeId.ExpandedText;

  //  var value = eventArgs.AttributeData.Value;

  //  var time = eventArgs.AttributeData.SourceTimestampLocal;

 

  //  Console.WriteLine( $"Value change: {recordId} - {value} - {time}" );

  // }

  //}

 

  private static void Client_EventNotification( object sender, EasyUAEventNotificationEventArgs e )

  {

      if (e.EventData == null)

        return;

 

      var Time = Convert.ToDateTime( e.EventData.FieldResults[UABaseEventObject.Operands.Time].Value ).ToLocalTime();

   var EventText = e.EventData.FieldResults[UABaseEventObject.Operands.Message].Value.ToString();

   Console.WriteLine( $"Event: {EventText} - {Time}" );

  }

 }

}

 

 

I do not know what causes the difference between your and my observation, but I suggest you retest your scenario(s) with our current version (2022.), to see if that does not resolve the problem for you. There was no fix made into this version for such memory leak, but there was a large amount of other changes, and therefore it is possible that there was an "incidental" fix.

If you still get the memory leak problem with version 2022.2, then make sure you provide me with truly reproducible scenario.

Best regards

Thank you for the details. I will now proceed with attempts to reproduce.

Note that I will be using the recent released version (2022.2), which is newer than you could have used. There was no specific fix for a memory leak in this version, but we do not officially support older versions, and if any code change results from the investigation, it would only be made to the newest version anyway.

Regards