Unstable Deployment on cRIO

[LukasKöpfli]

Hello

I have a problem with deploying an application to a RT cRIO 9038.

It is really unstable and unreliable. Sometimes it works, and sometimes it doesn't (without changing anything in the project).

I try to start the application with the run button from my Computer with Win10. The deployment runs throw without errors (mostly), but after the deployment finished successful, the run button don't get black. The application has not started at all.

Restarting the Target helps sometimes. Restart Labview helps sometimes. It is not reproducible.

Sometimes the deployment completes with errors in an arbitrary VI. If i open this one, there's no error. Try it again and the error is gone or in another VI.

Does anybody have the same problems with deploying to a cRIO?

We are a couple of developers with the same issue.

We're using LabVIEW 2016, with OOP.

The worst thing is, that it's not reproducible...

Best regards

Lukas

[MichaelBalzer]

---

@LukasKöpfli wrote:

...with OOP.

---

I find this to be a source of bizarre, non-reproducible deployment failures with cRIOs. I've had similar cases where a method of class claims it is broken because some other VI in the class is bad (but of course LabVIEW doesn't say which VI), only to close and reopen LabVIEW and everything is working again. Are your class hierarchies very complex?

You might try clearing compiled object cache and doing a mass compile, but I found that didn't help in my case. Anecdotally I do seem to have more issues with classes when the "Separate compiled code from source file" option in the VI is checked, but haven't done any in depth testing to verify the claim.

[LukasKöpfli]

Yes it is bizarre. We lose so much time with non-reproducible problems...

We do have same issues like you with broken classes but without any reasons and descriptions.

That's hard to say. It is not that complex. I think other languages e.g. C++ can handle this easily.

Class Hierarchy

 But we do have same issues with smaller projects.

Clearing object cache helps sometimes, but not reproducible.

I find out a new procedure to deploy mostly to the target.

If i'm restarting the target before deploying, the deployment runs mostly successful. BUT attention!!

I do the restarting with the internet explorer and the web interface from the cRIO. This procedure doesn't help if i'm restarting the cRIO from LabVIEW. It seems to be a difference between restarting from web interface and LabVIEW...