W.,

I do not have solution but have some something that hopefully can be helpful.

1. If the “write not complete” errors are not genuine, i.e. if they are not caused by the server taking very long to complete the Write, they would mean a bug somewhere – either in the server, or in the component.
2. Looking at the code, it seems that you /the customer have tried to use OpcAnalyzer for diagnostics. If so, have you verified that the server returns in time from the synchronous Write, or send a “write callback” callback in time in case of async Write?
3. Just a note: besides setting “desired method” to synchronous, you may consider setting “allowed method – synchronous” to False (I cannot easily figure out the precise property name right now).
4. The “hold period after topic write” should keep the item in SOME OPC group for that time, yes. It does not necessarily mean it has to keep it in the same group all the time. With “all defaults” (where Automatic update rate after topic write = Infinite) and just Writes to that item (no Reads), however, it should keep it in the same group. If it’s not doing it, I need to look at it, but cannot do so this week.

Zbynek Zahradnik

From: W.
Sent: Tuesday, February 15, 2011 4:02 PM
To: Zbynek Zahradnik
Subject: Write Not Complete Errors

Zbynek,

I have a customer who is getting WriteNotComplete errors with version 5.4.35 of OPC Data.NET. This Write error is causing the OPC Server to not let go of the connection from the client. So I am trying to figure out how to remove this error.

They have set the Timeout to 10 seconds which reduces the error but does not eliminate it. They have also set the Read method to Synchronous instead of Asynchronous. I have attached the code for you to review. This is the smallest application that reproduces the problem behavior.

Another point I noticed while on their machine is that the HoldPeriods for the WriteTopic do not appear to function the way I would expect. We set the HoldPeriods for the WriteTopic to 10000 (the write is on a 5 second timer). Each time the write occurred the group was added and then removed. I do not believe the timeout is caused by this additional overhead but I was trying to eliminate it as a potential problem. I tested here and see the same exact behavior with the TOP Server when writing. Does the WriteTopic hold period ensure the tag stays connected instead of removed and re-added? Or am I misunderstanding this property?

What I am looking for from you is any idea of how to remove the WriteNotComplete errors. Please let me know if you have any ideas.

Thanks,

W.

[attachment removed for privacy]