I am running HA in Container and today during an OS update I had a catastrophic boot failure on the NUC PC I was using.
I fired up another NUC PC I had laying around and put Debian 11 on it and copied over my backed up HA config files and all of my other support docker containers (mariadb, esphome, zwavejs2mqtt, etc)
Now when I start HA back up the CPU use is hovering between 65% - 100%.
I fired up another basic HA Container with a default config except I added system monitor and as soon as I stop the original HA Container (changing nothing else) the CPU load drops immediately to around 10%.
Everything that used to work on the old system still works on the new system except that things are fairly slow due to the CPU load.
any suggestions on what I can do to see where the CPU load is being eaten up?
I’ve disabled every integration that I could thru the UI (including cameras) and it made only a very small difference in load.
I also disabled my packages folders and it also only made a very small difference.
I guess it could be the CPU because it’s definitely less powerful than the other one but I wouldn’t think that even an underpowered NUC would be less capable than a RPi3.
Maybe that’s what it is tho.
Still trying to recover from the boot failure on the old machine before I take the drastic step of re-installing the OS.
You wouldn’t happen to know how to recover from a “grub_is_locked not found” error would you?
My bet would be on some database updates, that are run by the HA update. If something changes there, it uses a lot of CPU power to rearrange the DB.
What I’d def. try before I start with a new and fresh install, turn on debug logging and check what is causing the high CPU load. See the logger integration on how to do that. I tend to overlook that most of the time, but it gives some good insight, where exactly it gets stuck or what it is doing.
If I’m guessing right, letting it run for a few hours without interruption should cure it as well.
But let’s be honest, if you need it fast, setup HAos and restore your backup. But there are no guarantees!