teststand 3.5 execution crash

[danis_rom]

Hello,

i have the following situation: I start a sequence using batch model in Teststand 3.5 Sequence Editor (with 6 sockets for example). In MainSequence I start two new executions and then I wait after results.

After looping the sequence for a while (less than 10 minutes) SeqEdit.exe will crashes with following message: "Failed to create empty document".

This behaviour I testes on three systems, 2 NI systems and one time on my Dell PC.

My test sequence and some screenshots after crash you will find it in attachement.

I Use Teststand 3.5 (3.5.0.725)

I found some work arround to avoid this error, but I'm not sure if this is always working: in specify module->settings I checked in this option: Initially hidden and Disable Tracing. After that my sequence got no crashes anymore. I let it looping over 30000 times in weekend. My application should run for six months without breakdown 24/7, and the number of loops could reach 30.000 or more.

Thank you for your support.

[Scott_Richardson · Accepted solution]

danis_rom -

I ran your example under TestStand 3.5 after I unchecked the option to hide and disable tracing on the sequence call steps and I noticed in Task Manager that the GDI objects used by the TestStand sequence editor process increased with every iteration. I suspect that the TestStand sequence editor in 3.5 had some memory leaks associated with creating new execution windows. Running this in a OI will likely not have the same issue. However, to continue using TestStand 3.5 sequence editor you will have to keep the spawned executions hidden (easy), cache the executions per socket and reuse/restart them (hard), or you could call the sequences using a thread within the current execution instead of using a separate execution (easy). If you moved to TestStand 4.0 or later, it is likely that these types of memory leaks have been fixed.

[danis_rom]

Hi Scott,

thank you for checking my issue. I already took one of the easy way. I actually hide the execution, so no other windows will be spawn every time I call my new thread.

Have a nice day,

Daniel

[Scott_Richardson]

Daniel -

Also, I noticed that your sequence file had a CommonResult type with a version of 3.5.0.1. It did not appear to me that the type had anything different than NI 3.1.0.100 version of the type. Keep in mind that whoever opens the sequence files that contain this type, their copy of the type will be updated to your version. TestStand 4.1.1 has additional safeguards to prevent the propagation of this type into the type palettes. You could take a look at the following posting to learn more about this: http://forums.ni.com/ni/board/message?board.id=330&message.id=14125.