Revolution Pi Crash

volker · Post by **volker** » 26 Mar 2018, 22:22

55°C AMBIENT!!!! That means the temperature of your cabinet or room - are you living in the desert?
The CPU temperature will raise up way over 55°C. IF you really stress the system it will cross the 80°C border when the firmware is reducing the clock frequency in 100 MHz steps in order to gain stability. Reduction is not a fault condition but a Raspi feature and normally you should not really experience much changes in your application. When the system reaches a more critical temp stage the frequency is reduced down to 600 MHz. You will experience such a condition if you stress all 4 cores of the CPU with 100% pure calculations at 55°C ambient. But this is really not a typical situation. At room temperature you might experience a slight frequency reduction if all 4 cores are stressed 100%. This depends on the modules you have attached at the Revpi Side and their power dissipation (a gateway stays cool but AIO or DO with high output load will heat up and work like a toaster for your RevPi Core.

Our experts will have a look on your descriptions so far and might contact you by PM if we have any suggestions...

mezz · Post by **mezz** » 27 Mar 2018, 11:28

Hi Volker,

I don't know if your comment was directed to me but I am not worried about the 71 °C temperature itself, I just noticed that while the normal running temperature is around 50 °C after a crash it rises above 70°C: so a likely indication that something hangs (like a loop wasting CPU) and causes the crash and the system does not simply shut down.

This morning the revpid3 crashed again, this is the recipe, pass me the irony:

- revpi3
- sudo apt update, sudo apt upgrade
- several login session opening and closing, I used: 1 ssh console, 1 screen within ssh, 1 screen session was previously detached, 2 Filezilla sessions open

This time journalling storage was enabled and I have attached all the logs from the previous boot (the crash) and a new boot.

By the way, in the apt logs I haven't seen any kernel updates unless I am wrong.
Anyway, I am now going to flash the system back to the factory state and check again, in case of fault I am going to send the revpi3 back and you can test it up too. I'll keep you posted.

Any instructions on how to send it back to you directly to avoid waiting a long time to have a replacement? (delivery type etc.) You can contact me by PM of course.

Thank you, regards

Enrico

volker · Post by **volker** » 27 Mar 2018, 11:54

No, my temperature comment was a comment to Frank's worrying about 55 °C.
Yes if your temp raises up over 70°C my experience is that there is a very tight loop extensively consuming CPU time. It sounds very likely that your "crash" is instead a bug trapping the CPU in an endless loop so the system is no longer responsive. I've had some sw crashes of this kind during experimenting with all kind of applications but never with the pure os running in idle or running the standard system services...
please let our experts have a look first on your logs and then we will contact you by PM to tell you how and where to send your system. Where are you located?

mezz · Post by **mezz** » 27 Mar 2018, 12:35

OK thanks, I am based in Italy, Venice area.

Enrico

f.fallon · Post by **f.fallon** » 06 Apr 2018, 16:32

volker wrote: ↑27 Mar 2018, 11:54 No, my temperature comment was a comment to Frank's worrying about 55 °C.
Yes if your temp raises up over 70°C my experience is that there is a very tight loop extensively consuming CPU time. It sounds very likely that your "crash" is instead a bug trapping the CPU in an endless loop so the system is no longer responsive. I've had some sw crashes of this kind during experimenting with all kind of applications but never with the pure os running in idle or running the standard system services...
please let our experts have a look first on your logs and then we will contact you by PM to tell you how and where to send your system. Where are you located?

Sorry my misunderstanding, I tested with a stress program to make the pi undergo a thermal shutdown which was past 70 deg. I only thought it might be a problem as it was in a contained enclosure with no active cooling. I know now that is not the problem. I had it crash again today just mounted externally on a dinn rail. It seems to be a kernal crash with no way to recover from only if I reboot the device by power cycling it.

https://imgur.com/a/LWm00

It booted around 09:10:02 and crashed around 10:39:01

johnjsb · Post by **johnjsb** » 10 Apr 2018, 17:31

I too have the exact same problems. Crashing with no programs running just ssh terminal. Hope we can figure this out I have another revpi1 and revpi3 that seem ok. Im tempted to clone it to this one to see if it still crashes.

volker · Post by **volker** » 10 Apr 2018, 17:45

Sorry to hear you've trouble John. But "with no program" is not enough information to give any help. Is this a fresh image out of factory or have you installed things, configured things or done an apt upgrade? How do you mount the system? What is the power supply? What about functional earth (FE)? What are the symptoms of what you call "crashing"? No ssh? Any error LEDs? How about HDMI output? Or are there modules connected to PiBridge?
You see... tons of questions because there might be tons of reasons for the system to misbehave. So please explain what you are doing and what you are expecting the system to do and what the system is doing (or not doing).

johnjsb · Post by **johnjsb** » 11 Apr 2018, 16:32

Yes fresh out of the box turn it on. Lock up in terminal and all red lights on modules (i have 3 AIO modules and 2 DIO modules) at this time i did not have teamviewer or anything enabled. Hdmi being used shows same no interaction just locked up key press doesn't work just frozen. Using lab grade power supply with earth gnd works great with other units i have and is way cleaner than whats in the field so power is not it. So i upgraded everything still same. There are no programs running just boot and sit there seems to last about 2-4 hours. I did enable teamviewer (was not able to login for some reason) at some point not sure when but everything is disabled now left it running all night seems to be still up and running, will see if that has any effect. maybe upgrade and nothing enabled is the cure.

volker · Post by **volker** » 11 Apr 2018, 17:13

If you activated a service which is running into a fault you might get caught into a extraordinary log-file size due to some massive error logging. The emmc just has 1 Gb free space and once logfiles do occupy most of that space the system does run into weird behavior.
Hope you have read the documentation about TV RevPi. You do need a TV account with TV client version 12 better 13. It is only running on Windows Systems, not Android or Linux versions. Once you like to get this running and do have trouble please post in TV Subforum.

johnjsb · Post by **johnjsb** » 11 Apr 2018, 20:14

no huge log files. teamviewer on windows tried 12 and 13, i have account and all works fine at home and to work with other pc's. used teamviewer-info got credentials still didn't work.