Built Application crashes on deployed machine.

[mothra]

I'm having a problem with both LabView 2009 and 2010.

I've built an application that polls some deployed compact FieldPoints and compact Rios, selects specific channels, and displays a boolean based on if this value exceeds a range.

Reading approximately 40 channels across 1 cFPs and 4 cRios.

The application runs great in LabView on my development machine. 

I've installed the LV 2010 Run-Time engine as well as the entire 2010 Driver disk on two (2) separate computers. 

I build my application without any error and run my standalone executable on both of these machines.

On both machines, the application will run for a variable amount of time (10 min to an hour), before the app crashes. 

I've placed Error wiring everywhere that I can think of that makes sense, but the front panel doesn't display any error codes when the application crashes.

Is there some sort of dependency that I've missed?

[Cory_K]

If it was missing a dependency, I don't think the code would even run in the first place.

If I had to take a guess here, it sounds like you have a memory leak somewhere in your code.

Is there a possibility of this?

[mothra]

That doesn't make sense to me.

I just checked and the application crashes once at 64 MB Physical Memory, and then again at 80 MB.

This deployment machine has 4 GB of memory available.

And that doesn't explain why the program does not crash on the development machine which only has 500 MB of memory.

While running inside LabView, the the combined peak physical memory is 352 MB.

[Tejinder_G.]

This may be a little more difficult to narrow down, since there is no error being produced.  Have we ensured that if we connect error wires, we are propogating that throughout the entire application?

What if we step through the code using breakpoints, single-stepping, etc.  That might give you an indication of where the problem starts. 

[mothra]

I don't see how breakpoints will help me since it doesn't crash on the development machine.

I've wired up a few error terminals where it seems to make sense. I've caught a few errors from the cFP, neither of which crash the application though.

[Tejinder_G.]

Are you able to identify any differences between your development machine and your deployment machines?  If I understand correctly, when you run both the source code or the executable in the development machine, it runs fine.  But when running the executable on your target machine, you get this crash. Is that right?  Are there any software version differences?

[H_SH]

Sampling information is also valid to look in as if u try to take more points without considering proper syncronisation in base machine and field units system can crash ..

[mothra]

Tejinder G.,

That's exactly correct.

I'm using LabView 2010 on the Dev machine, and the 2010 Run-Time Engine on the target machine.

I used the same driver media on both machines.

Are there multiple versions of the 2010 Run-Time Engine? 

[mothra]

H S,

I'm new to LabView and I'm not sure I understand what you mean. Can you elaborate?

[mothra]

Update:

I've written a new program for a new test we're performing.

This application is of a significantly smaller scale, and lacks any sensitive information so I can share it with this forum.

It currently connects to a single channel on 1 cFP, but will be later expanded to read some additional channels probably on a cRio. (My other program communicated with 4 cRios and 2 cFPs. )

I'm getting the same problem; works on my dev machine but not my target machine.

I have identified a specific piece of code that seems to cause the application to crash immediately.

I'm using Data Sockets Read and Write to test that the network connection is active.

(If anyone has a better way to do this, TCP/IP Ping for example please share)

I'm posting my application for readers to scrutinize. 

Suggestions, criticism, and advice are all welcome!