Hi All,
I am a newbie to this.
After update to 0.60.1 Home Assistant not starting up. Its seems Home Assistant service gets stuck. But console never comes back up. Seems gets stuck on somewhere during startup. How shall I check what is going wrong?
There is nothing to indicate what went wrong. Can I go back to previous version?
Shekhar
Anything in the log? From which version did you update, did you check the breaking changes in the release notes?
Yes, I did check breaking changes. Nothing that should affect my configuration, I have none of these components set up. Nothing shows up in logs…
This is latest logs
2018-01-08 06:07:13 WARNING (MainThread) [homeassistant.setup] Setup of recorder is taking over 10 seconds.
2018-01-08 06:07:40 WARNING (Recorder) [homeassistant.components.recorder] Ended unfinished session (id=88 from 2018-01-08 13:58:02.939058)
2018-01-08 06:08:14 WARNING (MainThread) [homeassistant.setup] Setup of sensor is taking over 10 seconds.
2018-01-08 06:08:16 WARNING (MainThread) [homeassistant.setup] Setup of input_boolean is taking over 10 seconds.
2018-01-08 06:08:17 WARNING (MainThread) [homeassistant.setup] Setup of alarm_control_panel is taking over 10 seconds.
2018-01-08 06:08:17 WARNING (MainThread) [homeassistant.setup] Setup of group is taking over 10 seconds.
Gets stuck at with either input_boolean or group or something else.
Strange, how long did you wait? Maybe changing the logger settings will give some indication what is going on.
I will try to get the detailed logs tonight. I did not see any red flag in the existing log. I waited for an hour or so.
Are you using device tracking? I just went through a full re-install to find out that commenting out device tracking was the solution.
Yes, I am using device tracking. Let me try commenting it out,
No luck, do not see any error in HA logs or supervisor logs. Seems like all the components are loaded but nothing shows up on port 8123. Will wait for another day before going to full re install. Let me know if any one has any clue.
What is your latest changes in HA before you cannot load the HA page. Remove that change would help.
If you didn’t change anything in HA. You just upgrade to 0.60.1. How did you upgrade you HA?
Initially I upgraded HA using upgrade component. Which failed to initialize HA. Then I tried curl command to downgrade to 0.60 and then upgrade again. It seems downgrade to 0.60 is also failing.
Are you using hassio or ? Can you access ha via ssh and try to see the configuration file. Normally, HA cannot load if there is something wrong with yaml file such as space and indent.
I am using hassio, I can access ssh, see that mqtt broker is also started, all the add ons are running but nothing comes up on 8123 port on browser.
I suggest you check the yaml file by using yamllint.com . You might found that because of the space and indent from the yaml make HA cannot load.
yaml looks good. Finally saw something interesting in supervisor log. Can anyone make sense of it? HA restarts after this. Interestingly, 172.30.32.1 is not the ip address anywhere on my network. Not sure where it is coming from.
18-01-12 03:47:36 ERROR (MainThread) [hassio.api.proxy] Client error on API stream request Cannot connect to host 172.30.32.1:8123 ssl:False [Connection refused].
18-01-12 03:47:36 ERROR (MainThread) [hassio.api.proxy] Client error on API states request Cannot connect to host 172.30.32.1:8123 ssl:False [Connection refused].
18-01-12 03:48:34 ERROR (MainThread) [hassio.tasks] Watchdog found a problem with Home-Assistant API!
18-01-12 03:48:35 INFO (SyncWorker_2) [hassio.docker.interface] Restart homeassistant/raspberrypi-homeassistant
I had the same issue. The watchdog checks every ten minutes to see if your dashboard is working, so that it can automatically reset your instance for you. Unfortunately, if it takes more than 10 minutes for your components to load and for the dashboard to be accessible, it gets stuck in restart loop and never gets out. I worked around this issue by commenting out things I didn’t need from my configuration, which sped it up just enough to get me in.
This is my second post, and I’m not sure how to add a feature request yet, but I think the watchdog should be set for, let say, 30 minutes after a restart of homeassistant, then revert back to a 10 minute cycle. this would avoid this issue with long load times.
I am trying to reduce the components to load on startup. But what is the permanent solution for this? I can get it running by removing components but that should not be ideal solution, right?
Also was wondering I do not have many components to load, I have seen screens with way more configurations than mine. Getting frustrated by this, I cannot add more components now and my Phillips Hue light has arrived and I am afraid to add it to my configuration.
You’re right, it’s not ideal at all. What version of RPi do you have? i’m on RPi 1B+, which explains why mine is so slow. here are a few things I did that helped:
I changed from a highly distributed configuration type setup and put as much as possible in the configuration.yaml file and stopped using packages. not entirely sure this made a difference.
I turned everything off that I didn’t really use. There wasn’t much, but it was enough to get me in just under ten minutes and prevent the watchdog from triggering a reboot.
For my setup, the biggest performance hit I’m getting is by using z-wave. This takes a long time compared to other components, and I haven’t figured out how to force it to use MySQL instead of SQLite. This for me is a must have. Also, while monitoring my logs during a reboot, it seems to take a few minutes before anything happens. not sure what this is about, perhaps the supervisor layer is busy doing stuff in the background that doesn’t get logged.
The fix, from my perspective, is to add a reboot condition to the watchdog, which is something the devs would have to do. In my case, upgrading to a RPi3 would definitely help too!
You are right. I am also on RPi 1B+, and Z wave devices take long. I will try MariaDB way but not sure how much that will help. Also RPi3 is another option. Let’s see, If I happen to run into Rpi3.
I have an RPi3 for another project, I will port my installation and see what the performance difference is. (huge I suspect)
after a brief time trying it our on a RPi 3, i’m never going back lol. the appdaemon issue went away and the system boots up so fast I can do updates to the config and reboot without ripping my face off!