Hello Zbynek,

I think there is a misunderstanding at this point...

Your explanation is quite clear, that is why we keep the OPC server running with some other OPC connection.

When we browse the OPC server again with BrowseBranches(), the server is already running with the Dynamic items as well. An OPC client is already connected.

The BrowseBranches() returns only the Static ones.

The second part in your answer does not work, that's why I made this post.

If I try to read a dynamic item, then also BrowseBranches() work, but I can't do this in the reality, I do not know which Items are there.

Hello Zbynek,
I've just added the ZIP again.

On Friday I also sent an email to you, to thank for the license, but it also came back (rejected).

Everything went fine, thanks for the info.

Regards,
Emilio

Hello, I think this would be the easiest way:

I can reproduce the problem with an OPC DA Server that I created today morning. The real one is too difficult to use (it requires more infrastructure) and the problem is reproduceable with the fake one.

You can do the following:

- Unzip the content into any folder.
- Execute RegServer.exe (as Administrator if under Windows 7), without parameters.
- Connect the server (the name is: Endress.RuntimeNewItems.Test.1) from any OPC DA Client.
- If you browse immediately, then you see one branch, called Static, with 1000 items.
- Keep the connection alive, for example monitor one item.
- Try now with QuickOpc -> BrowseBranches(ServerDescriptor) -> you see the only branche.
- After at least one minute, connect another OPC Client (or the same like before, if it supports refresh) and browse again.
- You see 2 branches, Satic and Dynamic. Dynamic has also 1000 items.
- Now do it again with QuickOpc -> BrowseBranches(ServerDescriptor) -> you see only the first branche!

I give here a complete example of this issue.
I start an OPC DA server with an OPC DA client.
Initially the server creates 1000 items under a node called "Static".
The node is visible in the clioent and also its 1000 items.
One minute later the server creates a new node called "Dynamic" and adds there other 1000 items.
If I start now a second OPC client then I see the two nodes and all together the 2000 items.
At this point I start our application, which uses OpcLabs components.

When I browse this server, the first "browse" call (the rest goes recursively):

var opc = new EasyDAClient { Isolated = false };
var serverDescriptor = new ServerDescriptor(serverName);
var branches = opc.BrowseBranches(serverDescriptor);

NEVER returns the second branch, the one that was created dynamically one minute after first server connection.

This happens in all logical cases:
1) I start our application AFTER any connected OPC client is already showing the two branches.
2) I start our application once, wait some minute and browse again.

The second branche, at least in this example, is NEVER coming.

I made the following observation:
If I browse a dynamic node, just to test, obviously the software cannot know about nodes before the first BrowseBranches(), with the following statements:

var nodeDescriptor = new DANodeDescriptor("Dynamic");
var nodeFilter = new DANodeFilter();
var nodeCollection = opc.BrowseNodes(serverDescriptor, nodeDescriptor, nodeFilter);

and THEN I call the BrowseBranches(), then all items are coming!!!
They also come the next time (within a short time) also without browsing the specific node first.

We need a solution at this point: how a first BrowseBranches() (or some other instruction) can retrieve the current OPC Address Space, at least the immediate branches under the root.