Edge Computing Division Users Forum

I have a number of remote systems based on the fit-PC2-D1G-C1600-W-H8, running Windows 7 Embedded Standard. These are used for data-logging and diagnostics of our product.

Recently, two (in completely different locations) have stopped working under similar circumstances. I was logged into the machines via Windows Remote Desktop, and I was running some code that was possibly CPU intensive, or at least USB-intensive. After a couple of hours of doing this, the systems appear to have shut down, and I can no longer reach them.

I have the hardware watchdog timer running, so a system lock-up should have caused the system to reboot. The code in question doesn't run automatically at startup, so if the code somehow made the system crash, the watchdog timer should have handled that.

The system obviously starts up automatically after a power outage, so that doesn't explain it, either.

A network issue is, of course, possible, and this is what I assumed happened after losing contact with the first system. After I lost contact with the second system under the same sort of circumstances, though, I'm wondering if this is more than a coincidence.

Is there anything that would cause the fit-PC2 to shut down and not automatically restart? Overheating, perhaps?

Both of these systems are in Scandinavia (one in Norway, the other in Finland), so I wouldn't expect them to be in a very hot environment. Is there any way to monitor the temperature of the fit-PC and/or any way to throttle the CPU if it starts getting too warm?

Thanks!

Thank you for providing detailed description of your issue. There is a several cases may lead the system to shutdown and not come back afterwards.

* The thermal issue is less likely in your case. To overheat a FitPC2 you have to put in closed unventilated environment, what is, as I understand, is not what we are talking about. Anyway you can monitor a temperature by using a freeware tool called "speedfan". Please note, that this is a third party tool, unrelated to Compulab. Use it on your own risk. System does throttle CPU automatically in case of overheating, you don't have to do nothing.

* A software shutdown will remain the system off, until the user will power it on back

* I know that this is very suspicious when two systems go off simultaneously, but any way I would check the network.

An update:

I was able to have someone locally reboot one of the machines (which was off).

From examining the system event logs, it appears the culprit was memory exhaustion. One of the processes consumed all available memory, and Windows windows eventually shut down.

To prevent this in the future, I'm going to write some software to automatically kill the offending process and/or reboot the system if memory gets low.

Thanks!

Thank you for your update