Why did PXI RT reboot itself?

[nate]

We just had a PXI-8145RT board reboot itself.  This is new to me and we have had this type application running on this board for 7 years now without an issue.  The error log reports "C:\projects\labview\mercury_fullbuild\source\manager\memory.cpp(638) : DAbort: ".

The same application is running on several other identical PXIs without an issue.  

I have attached the full report. 

Can someone please suggest what could have caused this and what actions need to be taken to prevent it from happening again.

 Thanks

[Kevin_S]

Hi nate,

What version of LabVIEW are you running?  Do you have a system that deals with file storage or transmission?  (It could be a potential memory leak that shows itself only after extensive use times.)

Your issue sounds very similar to this discussion forum: Fatal Internal Error : "memory.cpp", line 638 

Kevin S.

Applications Engineer

National Instruments

[nate]

Version 7.1.

We do not have any file storage other then when initializing.  But we do send information to the Host PC via TCP.

Is there something more that I can do to help track where this might be occurring?

[Ben]

Tools >>> real TIme Sytem Manage>>> Resource tab shows memeory usage.

In RT yau can not use more memory than there is physical memory on the mahcine. Exceeding that limit will result in a crash.

There is also a switch (a config setting SOMEWHERE) that forces the RT to re-boot if the network connection is lost.

Ben

[nate]

Is there a log that shows you have ran out of memory before it reboots, or a way to log it if it starts getting dangerously low?

Also, why would we only have this issues on this one controller.  We have hundreds running the same application, and never have had

and issue with them.

[Ben]

"Why" questions are easier answered after we know "what".

Wild Speculation:

If an I/O is out of range etc it could return an error code in its error cluster. A populated error cluster requires more memory than an un-populated cluster. So in that case hardware errors could make this instance require more memory.

The above is just an example.

There are functions built into the project taht let you manage the RT app.

Ben

[Ben]

---

Ben wrote:

...

There are functions built into the project taht let you manage the RT app.

Ben

---

Pop-up on the RT node and look for something that looks like system management functions. From that window you can monitor CPU and memory.

Ben

[nate]

I'm not sure that I gave you all the information about the application.

The PXI that this happened on is a built executable at a customer site.  They do not have the LabVIEW development environment to monitor the device.  It has not happened again, so it's not like it's a constant re-occurring problem that can be easily found. 

So my concearns are

1.  The unit rebooted itself.  This can't happen!  It causes some major issues.

2.  Is it in fact a memory issue that caused the reboot?  I guess no one from NI can answer this 😞

3.  If there is a memory leak, is there a way to log it when the application is a built executable.

[Claire_Reid]

Hey,

See if either of these two links provide any help:

Is There an Error Log File for my Real-Time Controller?

Checking Available Storage Space on a LabVIEW Real-Time Controller

Let me know if you can see the log file and can run the attached VI.

[Ben]

Claire,

Can you post a reference that explains how to interpret the log from the RT crash?

Ben