HA dies after several days

Installation Home Assistant OS
1

Hi all. My HA (running on rPi 3) seems to stop working after several days. I can’t SSH or Samba to the device and I have to hard reset it to get it to come back on. Sometimes I have to reset it twice, but I’m not sure if it fails to come on or just dies again.

I’ve tested the SD card it and it tests out fine. For a lark I’ve tried a new one anyway and no change.

I’d check the log files, but it seems that the log files are deleted every time I reboot the system so I don’t know what is happening leading up to the freeze.

Any thoughts on how I can troubleshoot this issue? Was working fine around .99 with the exact same add-ons.

Thanks for any assistance!

2

What power supply are you using for the Pi?

3

It’s a dedicated rPi power supply. 2.5a or whatever.

4

Hi,
Can you check if the pi is still connected to the LAN when it dies?
I am having issues with home assistant runnning on ESXi, but this might be related.
my problem: Hassio on esxi losing IP address - HA supervised
Mine just loses its local ip address, however, I can still console to it from ESXi. you can try connecting a monitor to it when it is crashed, check if it is working.

5

Hmm. Won’t be the easiest thing to connect a monitor to, but certainly worth a try. When it crashes again I’ll see about trying to get video out of it. I’ve only ever run HA headless.

Of note, I think that I’m still able to ping the pi when I lose access to HA which makes me think the IP address is still alive but I’ll double check that too.

6

Alright, was able to plug in video. Monitor indicates that it is getting no signal at all. Seems like the device is D.E.D. Dead when this occurs.

Trying to determine if the problem is software or hardware related. Does anyone have a good way to check to make sure the Pi is running at its correct performance levels?

Thanks for any additional help.

7

Does the Raspberry Pi still respond to ping? I’ve had similar issues and once it locked up the Pi is still responding to ping, and LEDs are lit, but no one is home.

In my very brief experience, Home Assistant and HassOS have very little persistent logging for these types of issues, and even when actively monitoring they show very little extra. If it’s anything like my issue it’s resulting in some memory exhaustion that will freeze your RPi in default configuration.

I’d suggest you start with activating SSH access to the HassOS host system / hypervisor.

From here, you can at least do some monitoring remotely. I run several SSH sessions to check “top” and stream the home-assistant.log output (since this file is overwritten when home assistant restarts).

Another suggestion is to add more swap space (has to be activated manually after each reboot).

From the host (HassOS):

dd if=/dev/zero of=/mnt/data/swapfile bs=1M count=4096
chmod 0600 swapfile
mkswap /mnt/data/swapfile

Then to activate the new swap (has to be done after every reboot):

swapon /mnt/data/swapfile

More info on my issue, and the bug report I filed:

2 Likes
8

Yes, still responding to ping and LED is on (at least the red one) but nothing else. And you are correct, no video output at all so I can’t see what is happening.

I’ll look into being able to log into the OS portion specifically. My understanding of Linux is limited but I’ll at least try out the additional swap to see if that helps at all. I also don’t much appreciate the log file being cleared at boot up, and would prefer to see the prior log saved as log.old or something… but that is a topic for a different day.

Glad to know I’m not the only one experiencing the problem. I’ll monitor your bug report.

9

I had similar problem and stopped several containers like influx Addon because too memory consumer
Now it seems better now