I have a functional global variable (FGV) that runs on my RT target (LV 9.0f2) and is set via SubVI Node Setup to "open front panel when loaded" and Execution under VI Properties for the FGV is set to "Run when opened", "Allow debugging", & "normal priority". The interesting thing that occurs happens when I have already opened the project it is contained in (but have not as yet opened any VIs) and then I establish a remote browser session with the already running real time startup application on the remote real time controller. Everything works as expected, I can see my front panel in the browser showing that the application is running. Once I begin to populate the controller's inputs with data, the data triggers a previously uncalled state of a subroutine that calls the FGV for the first time on the realtime system to start tracking the data.
It is at this point that the actual VI on my host machine opens and runs locally. I do have a web browser page available on the target machine that I can establish a session with but I have not opened it yet, and besides, this is the actual LOCAL VI that starts running. It even stops running when I initiate the remote stop capability of my realtime application from my browser session. The VI then remains open on my local machine.
This only happens when the project is open on the machine I have begun the remote browser session on. Is this a "feature" of LabVIEW? I haven't seen any problems as a result of this, but the "linking" behavior is a little disturbing because it seems like there is some background linkage happening here that might also cause other problems - what else is happening in the background, gremlins? (if you know what I mean)?
By the way, my platform is Windows XP SP3, the browser is Mozilla Firefox 3.5.5, the application on the real time controller is taking about 25% of both processors (PXI-8106) capability and the data is NI XNET (PXI-8512 / 2) CAN related.
I'm not sure that I understand what you are explaining. But when using RT with front panels there is a background link going on. If you look at this article, and take a look at the Front Panel Communication Protocol section it explains it a little.
And there can be some caveats created from this. But there is no reason to think that this will cause gremlins in your program.
If the User Interface Thread fails (power failure of the host PC) the RT side will still continue to run.
Is there something in particular that you are worried about.
What I am reporting on is pretty simple. I have a VI that is self starting and runs when called on the Realtime platform. While I am viewing the front panel of the calling VI using a remote web browser session, when the data that starts the VI on the remote platform begins to flow, the LOCAL VI on my laptop appears to open and appears to run LOCALLY. This only happens if I have the project open containing the build specification for the project AND I have requested control of the top level vi from the web browser session. I suspect that this is not expected behavior. I have not tested it, but the VI that pops up like I have described is the only VI I have placed in the build specification under the "always included" category of source files. Upon closer examination I am noticing that the VI that pops up does NOT stay open when the upper VI is stopped remotely.
I guess this isn't really a problem, just not quite sure why it happens.