0.40.1 high cpu and responsiveness

Hey Everyone,

I am currently running HA on a rpi3 and have had to downgrade to version 0.38.4 several times after trying to upgrade to the latest updates starting with 0.39. I was hoping 0.40.1 would resolve the high CPU usage and timer sync issues, but unfortunately the lowest my cpu gets is about 8-10% and usually jumps to at least 25% with any activity. I’m not sure if the latest updates are just too much for the pi or if it is just a problem with my setup. I am running zwave switches, lights, and multisensors which seem to cause much of the high CPU. Sometimes, my pi becomes unresponsive for minutes at a time (not taking any ssh commands). After downgrading back to 0.38.4 my cpu hovers around 1-2% again. I’ve already offloaded my database to my synology using their mariadb. Anyone else have these issues that were supposed to be resolved with the latest updates? Is it just too much for the pi to handle now?

I am a new user, but I am having a similar experience. For me, I am seeing high utilization on a single core on a raspberry pi 3 for the hass process. In particular, there is one thread that seems to be chewing up all of the CPU. Here is a backtrace:

(gdb) bt
#0  0x76da2968 in select () at ../sysdeps/unix/syscall-template.S:81
#1  0x6d8a00c0 in OpenZWave::SerialControllerImpl::Read() ()
   from /home/homeassistant/.homeassistant/deps/libopenzwave.cpython-34m.so
#2  0x6d8a0104 in OpenZWave::SerialControllerImpl::ReadThreadProc(OpenZWave::Event*) ()
   from /home/homeassistant/.homeassistant/deps/libopenzwave.cpython-34m.so
#3  0x6d89f9a8 in OpenZWave::ThreadImpl::Run() ()
   from /home/homeassistant/.homeassistant/deps/libopenzwave.cpython-34m.so
#4  0x6d89f9c4 in OpenZWave::ThreadImpl::ThreadProc(void*) ()
   from /home/homeassistant/.homeassistant/deps/libopenzwave.cpython-34m.so
#5  0x76f30e90 in start_thread (arg=0x6cbff460) at pthread_create.c:311
#6  0x76da9598 in ?? ()
    at ../ports/sysdeps/unix/sysv/linux/arm/nptl/../clone.S:92
   from /lib/arm-linux-gnueabihf/libc.so.6

I have not debugged more than this - let me try to do some digging today, or if anyone more seasoned sees something obvious in the above, your help would be appreciated.

I found that triggering zwave soft_reset will cause this 100% CPU utilization indefinitely, 100% of the time for me.

When having below actions in my HA, it pops to 100% cpu for 1 core… (this is with HA 0.55.1)

alias: Heal Z-Wave Network
initial_state: true
trigger:
  platform: time
  at: '4:30:00'
action:
  service: zwave.heal_network

image

I have a similar issue with above lines, the docs do say “some Z-Wave controllers can cause the Z-Wave network to hang. If you’re having issues with your Z-Wave network, try disabling this automation.”

Strangly enough i do still get feedback from the z-wave devices from the devices itself to HA, but i cannot do any calls from HA to z-wave devices anymore. So my automation all still run but they can’t trigger the zwave devices.

So how do you heal it every night then? Did anyone make a ticket for this on HA github?