Actor Framework Discussions

cancel
Showing results for 
Search instead for 
Did you mean: 

AF on cRIO : Unbound computing node and other pain points

Hi,

 

We use a lot the AF on our projects on cRIO (unpacked and packed versions) and we often encounter problems with the IDE, which make the experiment really disappointing. 

 

One of the biggest one is the fact that the AF breaks during the deployment (i think because of the simple error handler badly compiled) referring to a subvi "not supported in the current target" (like if the conditional disable structures were not working).
Magically, it's not always the case. For the same source code, just clear the compiled cache, restart the cRIO, restart LabVIEW, burn a candle, and it works.
This NEVER happens with the AF on desktop. I don't know what we are doing wrong and there may be best practices that helps avoiding that kind of behavior.

 

Anyway, i'd like to focus a specific error JohnG3k and we had, and for which it would be great to have some feedback from NI.
It is described in this post:
https://forums.ni.com/t5/Real-Time-Measurement-and/cRIO-Unbound-Computing-Node-and-Deployment-Challe...

 

Does anybody know what would be the cause of this?

0 Kudos
Message 1 of 4
(1,043 Views)

This doesn't directly relate to the other post you linked to but in my experience, for whatever reason, deploying to RT and building PPLs sometimes messes up the active state of the conditional disable diagrams. (e.g. The dialogs in error handling VIs become active and cause errors on RT.) AFDebug has conditional disable diagrams that might cause issues as well. Sorry I don't have any direct answers - I'm just tossing ideas out there. 

 

I know cbutcher has a lot experience with AF, PPLs and RT - maybe he'll see this and can provide a better response.

CLA CLED AF Guild
0 Kudos
Message 2 of 4
(1,002 Views)

Sorry, I've never bumped into this issue ("Unbound computing node"), although I have many times had the issue you describe regarding polymorphic VIs not being a possibility (although that doesn't seem to stop be successfully deploying them).

 

I have sometimes had to deploy twice (but not build twice), in order to get it to succeed.

 

As was mentioned, I use an Actor Framework PPL exclusively when deploying to RT (as in, I never use the lvlib version).

My solution for the Conditional Disable structures was to remove them all entirely (along with the accompanying debug code).

 

I'm now running into a brand new (for me) problem that's killing my builds too, but I don't think it's related to Actor Framework. It might be related to outdated PPLs - I've heard that they behave differently on RT than Desktop (w.r.t. downstream rebuilds) and so I'm now rerunning the build to ensure everything is built based on the newest dependencies (over the last few years, I've mostly managed to save time by only rebuilding changed libraries, but sometimes rebuilding downstream PPLs also seems helpful for RT).

Edit: not a feature of out of date PPLs - still happening with a new set. 😕

A description (showing it occurring 5 years ago for others!) is here: Build process hangs (but it seems different to what you're facing).

 

Sorry that I don't have anything to add on the "Unbound computing node" issue...


GCentral
0 Kudos
Message 3 of 4
(994 Views)

Not much useful info from my side, either. I searched the Bug database and only a couple of references to 'unbound computing node' came up, and they're all really, really old, and don't pertain to the issues described here.

0 Kudos
Message 4 of 4
(978 Views)