I’ve been running piaware on a Pi2 since early April without issue until a couple days ago. It’s been crashing anywhere from 4 to around 24 hours after resetting it. The really annoying thing is I’m currently 1100 miles from home so I’ve had to have someone go power cycle it for me. I’m 99.9% sure it’s not a power problem as I’m using the 2.5A supply from the CanaKit. The only USB device is the SDR as it’s hardwired to my LAN. My suspicion is heat but it hasn’t been any hotter at home in the last few days than the rest of the summer so I’m not sure if that’s it or not. I’m far from a linux expert so I’d appreciate any suggestions where to start looking to try and track down the issue.
I would suggest you to check the temperature, if you have access to ssh of RPi.
When logged in, if you run this command, it will give you the temperature of cpu.
/opt/vc/bin/vcgencmd measure_temp
*
Note: 80ºC is the maximum allowed for a Raspberry Pi.
*
Thanks. I got someone to cycle it again and it’s been between 57-59C so far. Might see if I can find a script to output the temp every so often and see if it hits 80 when it goes down.
edit: and it went down again. The temp never passed 59C so it’s not heat. Guess I’ll just have to mess with it on Wednesday when I get home. So much for my 6 month streak.
anything logged?
dmesg -T
Unfortunately there won’t be anyone who can reset it again until tomorrow afternoon. Once it comes back up again I’ll try to dump that log. What’s also annoying is for some reason I can’t SSH into it from here even though all my router mappings are correct so I have to remote into my machine at home then SSH into it.
2 LtKernelPanic
Install the heatsink on the processor. Look at the CPU usage. There may be spurious unnecessary processes or services. Try it at least temporarily through the user interface FA disable support MLAT if you have it enabled (reduce CPU usage). We’ll see if next to the antenna ADS-B has posted some transmitting equipment (with strong signal to RTL dongle RF input USB bus may freezing). It is unfortunate that in the user interface on the FA website no opportunity to see the daily history of the system information sent from your RPi (every five minutes piaware sends a message to server containing basic health information about the local machine such as the system clock, CPU temperature, CPU load, basic filesystem capacity and system uptime)
Thanks for all the suggestions. I should be able to have someone power cycle it in a couple hours.
MLAT is off since there are no other feeder stations near me and the orange bar on my stats page bugged me. I honestly can’t remember if I put the heatsink from the kit on or not. I’ll make sure it is when I get home. Maybe something has cooked itself as I have the Pi and SDR dongle in the garage to keep it out of the elements.
I went through two 2.5 amp cana power supplies. Both started causing my rpi problems. I got a new 2.4 power supply from best buy. Have had no issues last 2 months.
I had a similar problem, it was the RTL 820T dongle overheating … I guess it then did something nasty to the power and took the Pi down.
Switched the dongle (the old one was now dodgy) and now run it with the plastic cover off so it gets air movement over the chips.
My Pi repeatedly crashed. It was operating remote. The only way to bring it back was to recycle power. Recently I was able to return to the site. I discovered by trial and error that the ethernet hub it was connected to was responsible. Apparently, the ethernet interface cannot handle the collisions occurring at the hub in a very small percentage of packets. I’m guessing at this point. Feeding the Pi directly into the router solved the problem. It no longer crashes. The router itself contains a small 3 port hub. So feeding a hub into a hub may create an issue. This flies in the face of the way ethernet and hubs should work. Nevertheless it’s something to try if crashes are an issue. Joe K4AA
BTW Belt and suspenders. I rigged up a z-wave device to cycle power remotely thru ADT>
could do a cronjob
*/6 * * * * ping -c 1 xxx.xxx.xxx.xxx | grep "Unreachable" > /dev/null && sudo reboot
replace xxx.xxx.xxx.xxx with the IP address something that should reliably respond. Maybe the DNS server of ISP, or even the internal IP address of your router.
This says every 6 minutes - ping the address, if the reply includes the word “Unreachable” - reboot the Pi
Test by by replacing the sudo reboot by echo would reboot and running the command interactively.
This would not be of benefit. When the Pi crashes it is instantaneous. It can’t ping. It can’t reboot unless you power cycle it. There is no way to access the Pi via SSH etc. Perhaps in the case of some dump/faup failure with the shell still responsive might benefit from this technique. That usually isn’t the case. Joe K4AA
If a power cycle is really required, how about using this…
blog.ricardoarturocabral.com/201 … ng-on.html
(obj has mentioned something like this before)
Once again I appreciate all the suggestions. After getting home today I checked my setup and nothing felt super hot but I still brought it inside and set it back up in my office window. If it seems stable after a few days I’ll try putting it back out in the garage and see what happens.
2 LtKernelPanic
If it doesn’t work well - try remotely via GSM. Choose your connector plugs.
Just an update for anyone who cares. I suspect heat my have finally been getting to the pi because after I brought it back into the house it was rock solid for the last five days just as before. I put it back out in the garage tonight so we’ll see how it goes. If need be I’ll rig up some cooling for the pi and SDR.