TestStand can't find variable once in thousand Set calls, "Error -17306"

[~jiggawax~]

Hi,

I don't think the version matters.  I reproduced it in 4.1.1 as well as 4.0.  Here it is all zipped up for you.  Just open the sequence and add a boolean to your station globals called Exit_Test.  Then run it and tweak the value on the popup for the wait control.  change it to like .1 and you'll almost immediately see the error.

Regards,

[Manooch_H]

Bill and Jigg,

When I first open the Sequence File that you provided in the zip file, I notice that the Run VI Asynchronously step is not configured to Show the VI Front Panel, so I had to enable that option. I then noticed that when I run the sequence if I try to set the Delay control to .1, it actually changes it to zero because the Numeric Control is set with a Representation of U32 as opposed to Double.

Nevertheless, even after making those changes, I have been unable to reproduce the error in TestStand 4.1.1 and LabVIEW 8.6 (which I assume is the setup with which you have been able to reproduce the error).

As this may be a race condition, it is likely that I will not see this error as often as others might. Do you have Tracing Enabled or Disabled when you're able to reproduce the behavior? Are you running the code modules in the Development Environment or the Run-Time Engine? What kind of processors are you running?

Thanks for your assistance in attempting to identify the source of this error.

[Bill2@NGC]

The PC is a Core Due E6550 @ 2.33GHz with 2G RAM.  The Trace is disabled, and I'm running from the seqEditor.  I'm not sure what's the distinction between the DE & Run-Time Engine?

[Manooch_H]

Bill and Jiggawax,

After looking into this issue further, I found that the particular sequence provided did not have result recording disabled for the steps executing in the While loop in the sequence file. This was leading to Out of Memory errors and other strange behavior. After disabling result recording, I was able to run the sequence for 18+ hours with a delay of .01 and .001 without any problems. Did you two have result recording enabled or disabled?

By default, result recording is enabled for every single step in a sequence. When executing several thousand iterations of a loop, it's generally recommended to disable result recording for the steps within the loop. I believe the use of Notifications was the correct route to take in this situation and it seems to have solved the problem.

Message Edited by Manooch_H on 07-14-2009 10:23 AM