Core update to 0.108 keeps failing

Hi,
I had this issue in the past but it was solved with ha su repair. Now, despite having tried several times in the past days, and even upgraded the supervisor from 214 to 216 in between attempts, the update keeps failing with this messages in the log

20-04-12 10:28:50 ERROR (SyncWorker_7) [supervisor.docker.interface] Can't install homeassistant/raspberrypi3-homeassistant:0.108.3 -> 404 Client Error: Not Found ("no such image: homeassistant/raspberrypi3-homeassistant:0.108.3: No such image: homeassistant/raspberrypi3-homeassistant:0.108.3").
20-04-12 10:28:50 WARNING (MainThread) [supervisor.homeassistant] Update Home Assistant image fails

Any clue?

Most common reason: out of disk space.

Yeah I found this suggestion on previous threads, but it doesn’t look full.
Honestly I have a few backups, I wouldn’t know how to free up (significantly) more space:

➜  ~ df
Filesystem           1K-blocks      Used Available Use% Mounted on
overlay                6962668   5401640   1189224  82% /
tmpfs                    65536         0     65536   0% /dev
tmpfs                   473080         0    473080   0% /sys/fs/cgroup
shm                      65536        68     65468   0% /dev/shm
/dev/root                96000     96000         0 100% /sbin/docker-init
/dev/mmcblk0p8         6962668   5401640   1189224  82% /config
/dev/mmcblk0p8         6962668   5401640   1189224  82% /ssl
/dev/mmcblk0p8         6962668   5401640   1189224  82% /addons
/dev/mmcblk0p8         6962668   5401640   1189224  82% /share
/dev/mmcblk0p8         6962668   5401640   1189224  82% /backup
/dev/mmcblk0p8         6962668   5401640   1189224  82% /data
/dev/mmcblk0p8         6962668   5401640   1189224  82% /etc/asound.conf
/dev/mmcblk0p8         6962668   5401640   1189224  82% /run/audio
tmpfs                   473080       844    472236   0% /run/dbus
/dev/mmcblk0p8         6962668   5401640   1189224  82% /etc/hostname
/dev/mmcblk0p8         6962668   5401640   1189224  82% /etc/resolv.conf
/dev/mmcblk0p8         6962668   5401640   1189224  82% /etc/hosts
/dev/mmcblk0p8         6962668   5401640   1189224  82% /etc/pulse/client.conf
tmpfs                   473080         0    473080   0% /proc/asound
tmpfs                    65536         0     65536   0% /proc/keys
tmpfs                    65536         0     65536   0% /proc/latency_stats
tmpfs                    65536         0     65536   0% /proc/timer_list
tmpfs                    65536         0     65536   0% /proc/sched_debug
tmpfs                   473080         0    473080   0% /sys/firmware

I have the same issue. Update to 108 went fine, 108.3 wont update. Also raspberry3. Plenty of space available.

Same problem here, trying to update from 107.7. No error in logs and the system restarts after clicking update but stuck on 107.7 when it starts up. Have tried rebooting as well

Yep I have this issue as well and I have loads of disk space available.
Not sure what’s going on as I cannot see any errors in my log files but I cannot upgrade to 0.108.3 from 0.107.7. I can upgrade the addons to new versions as well as HACS addons and I can also ping from the base OS Ubuntu 18.04. I also used to use AdGuard, but that’s no longer installed.

After a comple restart of the raspberry I was able to complete the update to 0.108.3

Okay, i got it working. Downloaded add on to check config against new version. Got a weird error first not really specifying what was wrong. Then i removed some unifi: lines in the config that was now deprecated. Tried again, this time it reported that something was wrong with my Amazon Rekognition (custom component via HACS). Removed those lines in the config and then it passed. Updated and it went perfectly. Reinstalled Amazon Rekognition via HACS and added it in config again. Rebooted and it worked and config check ok!

Thanks for your help. I did as you did, however the log for the New Version Checker gave me a load of stuff but nothing that helped. I checked it I had the Unifi lines in my config files as I knew I had added it via the integrations option in the past. It was there so I removed it from the config file and the update worked.
Thank you for your nudge in the right direction.

I guess I was wrong on the disk space, I deleted 200Mb of snapshots and the update went smooth.

You have to remember: a new image is pulled before the old one is deleted. So during the moment of the update 2 images exists on the (SD, SSD, HDD)