For a quick and temporary fix yesterday I downgrade supervisor to 2021.01.7. Be aware to create a snapshot first, some of my Add-ons where gone after doing this. After downgrade click on install they where back including the settings.
For downgrade use the Terminal & SSH add-on it’s still installing with the latest supervisor and execute: ha supervisor update --version 2021.01.7
Unfortunately overnight Home Assistant automatically updates supervisor back to 2021.02.6. So do not rebooting any addons then it keeps working.
Interesting, because I did the same, just went down to 2021.1.5, like described a couple of posts above. Automatic updates are off and downgrading worked fine. Just the error stayed and restarting or rebooting the whole Synology did not change anything.
Afterwards I even did one step further and also downgraded core as well, but this didn’t change the result either.
It also worked fine for me until I restarted an addon (DuckDNS by update). Have you tried restarting it? On the other hand, have you installed the addons from HA or by manually deploying docker?
I think I found a solution for that. ssh into your Synology.
Run the follow command to find the id for the supervisor container: docker container ls
The go into that folder: docker exec -it *id* bash
Then edit this file with: vi /usr/src/supervisor/supervisor/misc/tasks.py
Delete the lines: RUN_UPDATE _SUPERVISOR = 29100 self.sys.scheduler.register .....etc...... supervisor
How have you done to reinstall the addons? I just did that with DuckDNS but when starting it tells me that the configuration is missing. Have you stopped the docker from DSM first and then reinstalled? Or as?
you make nothing with docker
1- you must have "last good snapshot "
2- then install addon Terminal & SSH (not SSH & Web Terminal)
3- from Terminal & SSH execute: ha supervisor update --version 2021.01.7
4- then you restore only all addons from “last good snapshot”
when you not have “last good snapshot” have to make new config
everything should be OK … until:
the next supervisor update (based on info from “rdekruyf” happens every night)
and subsequent restart of the addon (is depend from you)
Edit: I’m sorry if this is duplicative of @blavak68 post above, but knowing my frustration with this issue I thought it might help someone
I’m back up to full functionality. What worked for me was:
Downgrading to Supervisor 2021.01.7 as described by @rdekruyf
Restoring a snapshot from 5 days ago (2021-02-05, I cannot recommend Hassio Google Drive Backup enough)
Upgrading to Supervisor ver. 2021.02.1 (I had migrated to Zwave JS and needed a 2021.02 version)
I hope this proves useful for those of you also having trouble.
@anschein I am able to restart without the new version of Supervisor being installed, I just have the notification on the Dashboard tab of Supervisor that an update is available.
Edit: @anschein upon further review I don’t think I understand your question
i mean in the supervisor ver.2021.02.1,can you reboot the addons?i have some connection problem with my MQTT broker and need a new version can reboot the addon
Thx,and i tried every version SUP i can and found ver 2021.01.8 is work for me and safety reboot the addons…But i didn`t upgrade to the new version of supervisor.
So can i ask which addons reboot fine in 2021.02.1?
like many others I have problems with the new supervisor. The problem did not occur until I restarted my synology.
I noticed that the new supervisor released in February bind the /dev:/dev path into the new containers (only in new ones or in those that the supervisor decides to recreate)
This thing creates problems for me with these addons:
-mosquito
-grafana
-esphome
I tried to delete the bind (dev: dev) but if I modify the container from docker / portainer the supervisor no longer recognizes the addon (the addon starts in this case)
@rdekruyf
Thank you very much for the time you’ve spent to come up with this solution. I first removed those lines from the config file and then repeated downgrading the supervisor. It worked somehow. I needed to ‘reinstall’ all of my addons, but this was quickly done. I assume, it just set some paths and did not have to download anything.
After a couple of restarts and restoring snapshots, it actually worked. All of my devices are gone and I need to repair them all again, but at least this is possible now.
I’m absolutely disappointed from the SnapShot-feature. After the first two or three times restoring an older version, the system still showed default values in e.g. the z2m config files. I also tried different snapshots from different weeks, but just shortly before I wanted to give up, my old configuration was restored. I don’t know why, but this feature looks more promising than it actually is.
To avoid that in the future, I will switch to the VM version of HA. Unfortunately, my Synologys are all running on ext4 devices, so I first need to get at least one drive emptied and then changed to btrfs. Let’s see how this works. At least, in this case the whole VM might be taken as snapshot and reverting should always be possible.
Anyways, always updating in fear that something will break ‘the evil and unsupported system’ is ruining the fun.
I have also decided to start an installation of the VM version of HA (I already have a hdd in btrfs) but for now it is on standby again because I have gone from having a 15/20% CPU usage to 75/80% just by having that VM up.
There must be some other way … I don’t know if someone with more experience in docker can advise us on whether it is possible to install debian on a docker and install HA from there. Synology’s Virtual Machine Manager is rather inefficient.
