I’ve had this happen with many other updates also, so I don’t think it’s only this version update.
When attempting to update, HA freezes at some random point (not the same every attempt), so troubleshooting is very difficult. Logging is set to debug, but there are no errors, warnings are benign, and debug details at the log stopping point are never the same.
This hasn’t helped:
ha su repair / ha su restart
as you can see in the snapshot. I’ve also tried
ha core rebuild / ha core restart
same result.
I have to keep trying many many times to restart and eventually I get lucky and it starts up very quickly. It seems that perhaps some integration is looking for data that it cannot find and hangs, but I can determine which one.
I eventually had to reinstall HA (downloaded latest .vdi file) and recover from backup (2024.1.2). It took a few attempts to finally get HA working again, but most of the addons were not running and could not be started. HA restarts to get the addons working got me back to the same status as above. Still trying before I go thu the pain of another complete reinstall.
Apparently HA supervisor and core is running, but integrations and addons are not. See below, supervisor is indicating a healthy status and banner command returns fine.
Thanks Mitch, I have all my apps working. I have no issues there. My issue is My automation keeps rebooting my system. I ll have to turn it off until a resolve is found. I have attpted to install the following versions with no sucess.
Hmmm, I can’t even access any of my addons/integrations because HA gui isn’t fully loaded (e.g. “Starting Sun, not everything will be available until it is finished”). Also, looks like my addons are actually started and stable (see below) but HA gui is still frozen.
It seems like the gui (particularly dashboards) and startup freezes if a critical integration is not fully loaded. I can’t figure out how to determine which is that particular integration since the logs are of little use.
Even if I did, how do you restart with only specific integrations/addons? And I wonder if rebuilding my entire HA platform integration by integration is the only way to find it (I think even trying this still might not isolate the problem).
For what it is worth but I had to switch back to 2023.12.4 all 2024 builds fail after a few hours than the Home Assistants hangs. Tried 2024.1 2 and 3 but all fail after some hours. 2023.12.4 keeps running.
Restoring backups don’t always work for me. I’ve tried every possible way to recover including my last, which is a full start-over.
A fresh install, of course, works fine. But as I start adding Add-ons and integrations, it becomes a hit/miss on successful restarts. I’ve tried restarting after each change, then roll back that change if I get a failed restart only to get another failed restart.
I’ve been able to get back all of my configuration with the latest versions, but I have to attempt restart multiple times. The only success I’ve had (so far repeatable) is to use
ha core restart
from a putty terminal (since the GUI freezes - no access to Developer tools/restart).
This usually fails on the first attempt, but if I ctrl^C just after entering this it will sometimes give me a clean restart.
Typical seqence is:
ha core restart (then gui freezes at some point)
ha su repair
ha su restart
ha banner (only to monitor supervisor restart)
ha core restart
As I originally mentioned, the logs are of no help. All Add-on’s / integrations work well when HA is up and running, its just getting past a restart. Something is happing during that process that causes it to freeze. Access and commands from the (external terminal) command line work fine.
It would be nice to have better feedback (like progress, errors, etc) notification and logging during the startup process and options to restart with selective Add-on’s and integrations. Would also be nice to ‘kill’ from the GUI and restart with these options.
this happened 1 hr ago: Matter server update crashed my ZHA (I have sky connect with multiprotocol) and then crashed the whole system- I am at my 10th hard reboot because I can only access the GUI for 10 mins at a time and then I lose connection
unbelievably bad piece of hardware this SkyConnect, I deeply regret buying it - not only it is unstable for zigbee, now it crashes the whole installation
Well I finally decided to strip my core.config_entries file (I know, danger-danger!, but what did I have to lose?) down to a fresh install which of course restarted with no problem. Then I added back integrations one at a time.
The good thing about this is everything else (db, files, entities, etc.) all stayed in tact, just not usable until the applicable integration was installed.
It took a little time, but far, far less than trying to restore anything from backup and waiting at each startup attempt only to find it had frozen again on me.
After each addition of an integration to the core.config_entries file, I restarted. Only a couple of times did it freeze. I backed tracked an integration or 2, then was able to continue. Finally, all integrations are installed, everything back running and all with the latest updates. No clearly offending integration found.
There is certainly a ghost in the software (I’m sure its conflicts with integrations) preventing it from cleanly restarting reliably.
Again, would be nice to have a kill switch when restart doesn’t complete, so you can disable integrations one at a time and get back running more quickly.
Found the problem (painful trial and error). It’s the Google Nest integration when you have more than one home configured. I’m guessing it conflicts with other Google integrations since it uses access tokens.
I have to delete the additional home integrations from core.config_entries every time I have to restart (but leaving my one original home). After restart, I have to manually add the additional homes back to the integration from the Integrations page.
