PeteAsking

Something has gone wrong with your local template. It has not pulled the latest template. Can you remove this broken one and also delete the local template from your /boot/config/plugins/dockerMan/templates-user directory so it is forced to download a new template?

AlphaTester: 11notes/unifi:10-unraid 10-unraid is an alpha tag/you are a tester. Thank you for testing, it is appreciated. There was an issue in the latest build. Please restore from backup as per your agreement to use this tag only if you make backups. You can also try downgrade the tag to the previous version you were running. If it starts please take a config backup, and then redeploy the latest tag available (unifi:10.1.85-unraid) and restore the configuration. If you are not expecting to have issues, please dont run an alpha channel in production. Many thanks, P

If you dont post any information, what version you are on, what tag you track, when it last updated, if you changed any settings etc I cant help you. This is not a guessing thread. YOU post useful information if you need help.

You already have a different docker using port 1900 as the error says.

Thanks. I forgot that unraid stops the container but for some reason never starts it again so after a restore you have to manually start the container after. Eventually we will work on a single docker container for the unifi os server as well so seems like we can hopefully find a solution that lasts forever.

@jumperalex In your docker config your extra parameters should be in this format: --memory=8G --tmpfs /unifi/log:exec,uid=99,gid=100 --tmpfs /unifi/run:exec,uid=99,gid=100 --tmpfs /tmp:exec,uid=99,gid=100 Most likely you have only this: --memory=8G --tmpfs /unifi/log:exec,uid=99,gid=100 --tmpfs /unifi/run:exec,uid=99,gid=100 --tmpfs /tmp:exec,uid=99,gid=100 needs to be added as the docker uses the /tmp directory to facilitate the restore.

Yes this is caused by a missing parameter in your docker config. I will go get the setting for you.

Then the last version you can run is 10.0.162 due to your processor age (17 years).

If your cpu doesnt support AVX then you can no longer upgrade. You should check if your cpu is that old.

You need to take a backup of your controller and then restore it to a newly deployed version running that latest version. See all the posts above. You cant upgrade. It must be redeployed and restored.

Interesting but one issue is if they change things it can break for unraid so we need to ensure they become aware of unraid if we use their image. I will ask 11notes if he maybe wants to build an image like this.

Sure I will take a look and see whats up

Unifi will eventually drop support. Nothing anyone will be able to do to stop that. No they want you to start using it so will be nagging incessantly from now on.

No your backup file will work. You must take a backup on the 10.0.162 docker (or any 9x docker for anyone with a 9x version) and then deploy a brand new docker that is 10.1.85 that has not been upgraded at all in any way. Then restore that file to the 10.1.85 docker. If you are able to ssh into the unraid box you could stop your old container, mv the folder to an archive location, then delete your orphaned docker and deploy a new one running 10.1.85 and do the restore. WIth a copy of the old files archived somewhere you can potentially go back if you wanted.

Everything updated, so long as everyone takes a backup and then restores to a brand new container for 10.1.85-unraid everyone will be fine. Anyone running the 10-unraid alpha tag will break their deployment and need to restore from a backup. Anyone manually changing tags will be fine so long as they take a backup file and restore to a new docker.

Ok I will fix up the docker container xml and have to stay with it myself then to help support it. I will update the first post today later with various instructions.

Fair. Give the new network controller a go then. To get onto the latest you need to take a backup of the current one and then deploy the new one to a new path and restore the config.

For people who want to continue with the network controller can you raise your hands? To use the network controller you must take a backup of your current controller on whatever version you are on. Then you must stop that docker/rename/delete/archive it as we will be using the backup file you have. Deploy a new version (Be very careful to check you are not using the same paths if you opted to keep your docker instead of deleting it as by default it will use the same locations on unraid and settings). EG: Expand "more settings" before clicking deploy and change /mnt/user/appdata/unifi-controller-reborn to a different path eg: /mnt/user/appdata/unifi-controller-rebornNEW Access web UI and choose restore from backup - use previously downloaded file. (Any issues here with restoring let me know as we fixed this very recently so if you try within the next 4 hours, that fix may not have propogated down to you, but I can manually tell you what to do). This seems to give you a docker network controller that works. At this point can I ask who would like to use the network controller over the VM method? Reason I ask is there is an advert on login to move to the new VM version form unifi so its pretty clear they want you to move. Pete

It was pretty basic under debian. I think the only commands were: sudo apt-get install podman wget https://fw-download.ubnt.com/data/unifi-os-server/1856-linux-x64-5.0.6-33f4990f-6c68-4e72-9d9c-477496c22450.6-x64 sudo chmod +x 1856-linux-x64-5.0.6-33f4990f-6c68-4e72-9d9c-477496c22450.6-x64 sudo ./1856-linux-x64-5.0.6-33f4990f-6c68-4e72-9d9c-477496c22450.6-x64 Then wait for it to install and then goto https://ip-of-vm:11443 (take a long time to start initially). Then update the network controller (threre is a section to do this in the web interface) Make sure docker is turned off so you dont have 2 controllers running at the same time. Then do a restore under the relevant location (also takes a long time). Then its up and running and adopts all your devices. I had 2 access points I had to reboot manually to get them to adopt but otherwise all worked. P

I think everyone should just deploy unifi os server to a vm. I just did it. Took like 30 minutes and its more full featured and better than the controller anyway.

Please do not update image to 10.1.85 as this breaks the controller. Deploying a new controller on 10.1.85 is working but cant seem to restore a backup file to it. My suggestion is to move to a vm with unifi-os-server at this stage still.

Interesting script but I would use the podman method personally. Unclear why they bare bones mongodb when it can be containerized.

For those following along a potential fix I dont understand is being discussed here: GitHubFailed to build 10.1.85 · Issue #34 · 11notes/docker-unifihttps://github.com/11notes/docker-unifi/actions/runs/21976591788 So while Im not promising it will be fixed it seems like there is a chance we will get another 2 years or so out of this docker if that works out. @jumperalex @fat @wayner @wgstarks @bmartino1

Yes its possible to do this yourself but not via unraid gui I guess which was the goal. Personally I wont be going that direction but you can of course.

I dont know how to run full docker compose on unraid. Seems like unraid is just single docker deployment. That link to 11notes unifi network application has no benefit over just using the one other people are using here already where you deploy a mongodb docker totally separate and then the network application after separately on unraid. There is already a forum post and videos how to do that on unraid.