LVOOP and RT Issues - Anyone seen these?

[kent.wedeking]

I'm seeing two potentially unrelated issues in my current application.  One is minor ... the other is (potentially) life threatening to my system ...

First, some background .... I'm in the middle of a large development project (Windows for the GUI, PXI-RT for ... well, everything else).  This is also a very heavy OOP project.  We are using LVClasses for as the messages being transferred between modules (individual paralell vis) and the TCP Communcation Engine, as well as for configuration and data file management.  From an architectural standpoint, this system is basically a "Hub and Spoke" model where there is a Hub that recieves and directs messages from all the spokes.

so issue #1 ... during development when I deploy my RT to my PXI (all I'm doing is clicking the run arrow) the front panels of certain vis open DURING the deployment process.  One of them does have the front panel set to "open when called" while I'm doing my debug.  The other is a the front panel of a dynamiclly spawned clone that I don't need or really want to see most of the time.  since they are opened during deployment and not when called, then don't close at the end of execution either.

and onto #2, this one is a problem ... fairly scary one.  When I deploy said RT system by clicking the "run" arrow I'll get to some point in the deployment process and I'll get a fatal error "F146640: Thread Stack Overflow."  I've done some looking into this, and I highly doubt this is related to the on KB article I could find (Real-Time Fatal Kernel Error 146640 at Boot Up) because this isn't happening on boot-up of the system, but rather on deployment.

That specific error says that if I have the serial console activated (right now I'm only using a monitor hooked up to the PXI DVI port) I can possibly download more error information, so I'll try to find a null-modem cable and see if I can't make that work.  If anyone has seen this or something like this I'd love to hear about it.  We are leaning in the direction that these issues are LVOOP related because we've never seen them before, and this "isn't our first rodeo" on an app like this.  I'm not pointing fingers ... I love LVOOP ... I just need it to work ...

[christian_w]

Are you already working with LV-RT 2010 SP1 ?

[kent.wedeking]

I'm at v10.0f4 ....

[christian_w]

I suggest you to update LabVIEW and the RT module to SP1 first.

[StephanieA]

If upgrading to SP1 doesn't work, I would like to take a look at the serial output information if you are able to obtain it. Feel free to create a service request and talk to one of our applications engineer about this issue since we will need a lot more information in order to properly debug and test the issue.

[kent.wedeking]

sounds excellent.  I probably should get an SR# made for this sooner rather than later ...  I'm in the process of installing the SP1 updates for LabVIEW and LVRT, and then I'll need to update software on my PXI ....

[kent.wedeking]

FYI ... upgrading to SP1 did not help.  I've tried doing multiple mass compiles of all my code to no avail.  It happens at random.  Some days some front panels will open ... i'll modify some code, and then front panels of other vis will open.

I solved my fatal kernel error (i think)  I had one or two references that were not being closed ... Once I made sure I was destroying all my refs my fatal kernel errors have gone away.

[StephanieA]

I'm glad you were able to find the cause of the fatal kernel error by closing your references. Did you ever create a service request for troubleshooting this issue? If so, do you have a reference number that I can look at?

Also, can you post your project that has the Front Panels opening upon deployment? I have personally never seen this type of behavioir occur and am curious to see if I can reproduce it. What is your host computer and what is your target? Do the changes that you make seem to be of the same type (moving a control or adding a VI of a certain type etc?).