Hi, I am pretty new to the Home Assistant (HA) community, a lot of this is very new to me coming from a six-year relationship with SmartThings, Rule Machine and Webcore. I am finding HA to have a very steep learning but fun learning curve. I am just lost on this.
Home Assistant Blue - HA Blue
PROCESSOR - 6-Core Amlogic S922X Processor (ARMv8-A), Quad-core Cortex-A73 @ 2.2Ghz, Dual-core Cortex-A53 @ 1.8GhZ
MEMORY - 4GB DDR4
STORAGE - 128GB eMMC Flash
Earlier this week I saw in the binary sensor that there were updates available. In the supervisor dashboard (Supervisor > Dashboard) it showed that there were three components to update
1. Home Assistant Core - Current Version core-2021.11.5
2. Supervisor - Current Version supervisor-2021.10.8
3. One other that I can not remember as HA did the update and the third box disappeared…. maybe updated the OS?
The system was backed up, the update on #03 (above) was performed, and the system restarted from Supervisor > System > Host. After the restart there was a new warning in the Supervisor > System > Supervisor area that my system was unhealthy.
Reading the forums etc. for the last three days I see that it is not an uncommon error but none of the fixes seem to work and some of the threads just… end.:
1. ~110 gb of free space, so AFAIK, not a problem
2. Added the open DNS sever JIC it was network related
3. Restarted each HA section individually
4. Cold restart (off 30 seconds / restart)
5. Shut down one hr. / restart
6. Tried to restore from the back up but this does not seem to do anything. When I press restore and the subsequent pop up to wipe and restore buttons the system pauses for a second, then nothing seems to have changed.
One other thing that may yield information to those that actually know what they are doing here, ever since the update (#03 above) the RAM in the HA Blue gets used and not released @ a prodigious rate. The system starts ~6% and within 15 or 20 minutes is 87% to 90% used until the system becomes nonresponsive and has to be power cycled to resuscitate the HA Blue. The longest the HA Blue has been up since the update is 60 minutes.
Any suggestions, pointers, links or ways to get 12 to 18 hours of my life back?
Apologies, for some reason this did not paste into my original post:
A manual update was attempted and this was displayed in a pop up and shows up in the supervisor log:“21-12-22 11:34:09 ERROR (MainThread) [supervisor.supervisor] Update of Supervisor failed: Can’t install ghcr.io/home-assistant/aarch64-hassio-supervisor:2021.12.2: 404 Client Error for http+docker://localhost/v1.41/images/ghcr.io/home-assistant/aarch64-hassio-supervisor:2021.12.2/json: Not Found (“no such image: ghcr.io/home-assistant/aarch64-hassio-supervisor:2021.12.2: No such image: ghcr.io/home-assistant/aarch64-hassio-supervisor:2021.12.2”)
21-12-22 11:34:09 CRITICAL (MainThread) [supervisor.core] Can’t update Supervisor! This will break some Add-ons or affect future versions of Home Assistant!”
I am not sure what changed but the RAM usage issue has ceased, the HA blue unit has been up for 24 hours RAM holding ~20% the same as previous. The “unhealthy installation” notice, inability to update supervisor or core to the latest versions and the 404 error still exists.
Does anyone know if pulling the container image (Working with the Container registry - GitHub Docs) will alleviate / ameliorate this issue? It is a bit beyond me currently but if there is a consensus that it might work, I guess it would be worth learning.
The RAM hole and the “unhealthy installation” notice, inability to update supervisor or core to the latest versions and the 404 may be unrelated. Thirteen hours after the last post the HA blue started not releasing RAM again. The only difference between the day everything worked the next day was a shower. I disabled the shower related automations and things worked a full day, then quit again. This time the difference was that the upper bathroom was used. What those things have in common is an HACS addon called “Average Sensor” (GitHub - Limych/ha-average: Average Sensor for Home Assistant).
After rem’ing out the average temperature (upper bathroom) and average humidity sensors (all bathrooms) the RAM usage instantly stopped and the HA Blue has been running fine for the last couple of days. It is possible that this Average addon may be the memory hole, as I am such a tyro I am unwilling to state such a thing as a fact, maybe this will help others with diagnostics.
After disabling the above sensors and almost every integration I am still unable to update the HA Blue supervisor or core versions.
RAM Hole: Another day of up time, I think rem’ing out those average sensors removed the RAM hole issue.
HA supervisor update: Sometime between my last post (AM 12/29/21) and 1600 hours, same day, the “unhealthy installation” went away of its own accord and the HA supervisor updated itself. No additional changes made on my part
Unfortunately…
HA Core: Now the HA Core is giving the same 404 error and can not be updated.
It seems that this 404 error (in this instance) is not a connection (quality / speed), DNS or HA space issue as if it were the HA Supervisor would not have resolved itself. As a scientist I am reluctant to “just leave it and let it resolve itself”, it needs a root cause analysis to prevent the issue or at least make it less of a mystery error. Does anyone have any ideas that can be pursued as a possible root cause?