I am running HA on a Dell MiniPC (actually a former Chromebox, converted to a dedicated HA machine with MrChromebox.tech firmware mods to kill the old Chrome OS and to allow any other OS (all flavors of Linux, Windows, and of course HA) to boot. About a year ago, I installed HA OS and it has been running flawlessly all that time…until the latest OS update (16.3) and Core (2026.1.0) update. A few days ago, I initiated an update of the OS from 16.2, and unfortunately updating was just not possible, and things were stuck on 16.2. Worst still, when the failed update completed and HA was again accessible on my browser, many of my more recent device and automation additions were simply missing, and I observed that HA was loading an older configuration file that did not include all devices added over the prior five or six days before attempting the update.
TL: Long story short, I saw in several posts here on the forum that because my “Slot A” had likely become corrupted (for reasons unknown), HA was locked into booting from only “Slot B”. I tried a number of CLI commands suggested by ChatGPT to make “Slot A” good, but none worked. Finally, I came across a message on the forum which suggested simply using the boot menu, and catching the “Slot A” /“Slot B” arrow key selector during the boot process…in other words, quickly forcing “Slot A” to be the boot slot…and it worked. Finally…The OS is now at 16.3, but I am still unable to update the Core beyond 2025.12.5. And indeed, any attempt to update the Core to 2026.1.0 results in just a continual string of Filesystem CRC errors which go on ad nauseum until force terminated (with a “ctrl-alt-del”), to interrupt and halt the failed update.
Further reading of other posts and several AI/Google searches on the subject of why Core 2026.01. does not work on some machines, resulted in a suggestion to run a CLI “core check”, which returns an “Error: Object of type DockerMount is not JSON serializable.”
A Google search of that Error, returns the following:
"Based on community reports, the Home Assistant Object of type DockerMount is not JSON serializable error in version 2026.1.0 is a known, tracked issue, particularly affecting user interface access.
Symptoms:
- After upgrading to 2026.1.0, the UI works briefly, then displays a 502 error.
- Automations continue to run in the background, but the web interface/frontend becomes inaccessible.
ha core checkor supervisor logs may show the “DockerMount” error.
Temporary Workarounds / Solutions (As of Jan 11, 2026):
- Reboot the Host: For many, a full reboot of the host machine fixes the issue temporarily. A simple core restart may not work.
- Downgrade: Users reported successfully returning to 2025.12.4 or 2025.12.5 to restore stability while waiting for a patch.
- Check Docker Version: Some related issues are triggered by Docker upgrades (e.g., to Docker 25), causing incompatibility with the supervisor"
Have others experienced this same problem of not being able to update HA OS from 16.2 to 16.3, and/or not being able to update Core from 2025.12.5 to 2026.1.0?
If so, is there a fix, or do we simply wait for another update that will resolve the issue(s)? Any assistance or insights would be greatly appreciated.
BTW, this is the first instance of HA system instability that I’ve experienced. It’s got me pretty concerned because over the course of the past year, I’ve really heavily invested time and effort in integrating and automating just a ton of devices in our home, and if I am looking at only more future instability that will make HA unreliable or worse, inoperable, I will need to look either at buying other hardware to run HA on, or look to alternative systems. And our current home setup is extensive: 106 Zibgee devices, 32 Wiz bulbs, 63 Lutron “Clear Connect” switches and Picos, 92 Yolink Lora devices (mostly leak and motion/door/window sensors), and a smattering of many other brands and devices (e.g., Aqara, Sonoff, Roku, Midea, Rachio, Soma, Somfy, Tuya, XSense, Nest Smoke Detectors, 15 Eufy cameras, LG smart washer & dryer, and a dozen or so Ikea devices (both Zigbee and Matter over Thread). That’s just a quick summary, but the point is that if HA becomes unstable, that’s going to be a real problem.
My first thought was to see if I should just reinstall HA OS on my Dell machine, to fix any Filesystem errors… but then my next thought was that perhaps the UEFI that MrChromebox uses may not be “standard” and that could cause issues with updating to 16.3 (at least one Google search suggested that 16.3 now conducts some sort of “verification” test and that if the UEFI isn’t plain vanilla Intel, it can create a block on “Slot A”… But now that 16.3 is loading (perhaps only through a “force” boot with some quick up-arrow action during the boot process to force HA to boot from “Slot A”), I’m not as concerned with that…as long as HA will now boot from Slot A as the default…I need to reboot again to see if that works.
My second thought is to simply buy another Mini PC, something like a Beelink or GMKT Nuc, reload a “clean” OS version, and simply restore my devices from a complete backup.
In any event, have others experienced issues with updating either the OS or Core? And if so, how have you resolved the issues? Or do we just need to wait for a patch to resolve issues?
Thanks for any help and assistance.