Yesterday evening, I accidentally migrated to Eel (in short, I thought the stable version would be available the 24th, so I change the version, saw that it was still the RC version… and forgot to change back to Dragonfish).
Anyway, I discovered this change this morning because 2 apps failed to be migrated in my system: Homepage and 2FAuth.
Homepage, it is actually ok, just that now the app does not have an defined URL, it tries to connect to “http://0.0.0.0:30054” instead of “http://xx.xx.xx.xx:30054”
Can anyone explain to me how to indicate to use my url?
BUT my big problem is with 2 FAuth.
Actually, it is not displayed in my list of Apps, it has been simply deleted. I still have my dataset but the containers and the settings just disappeared.
This is a known issue in the RC. In internal testing we saw that the button corrected itself after manually navigating to the correct URL or even refreshing the apps page, but some have said that didn’t work for them. It is fixed for the .0 release version.
Not sure about 2FAuth. Is there anything relevant in /var/log/app_migrations.log?
Thank you,
no, they were both Truenas, I installed both less than two weeks ago.
I noticed more & more curious things:
If I go in the “discover apps”, they are ALL indicated as installed, which is quite untrue.
I Installed another instance of 2FAuth, everything was fine, but at the end, it is again the same problem, the web UI points to an IP address http://0.0.0.0:30054 and I do not see what is the problem, and where to fix that.
Unfortunately, no.
The apps are not accessible, even if I type the expected url.
And for all my other apps that work perfectly (I forgot to mention but it is the small part, I have 13 apps installed, only 2 have problems), if I click on Web UI, it opens the correct URL.
Edit: I tried to install another app, and it works great (besides the problem with the web ui). So it seems my problem is really specific to those 2. Let’s see the bright side
So in fact, I did more tests with Homepage at least and the problem is with the Host Path.
To explain,
If I create an new instance with Host Path, I always get the same problem, cannot access, and got the same error copying config.
If it is with IxVolume, works like a charm. But as it is based on yaml files I have to modify, I find this quite annoying.
Seems really to me there is a problem with those 2 apps specifically, the new app I have installed later is with an Host Path and there is no problem.
And this kind of issue will be resolved when ElectricEel will be officially deployed, so in… 5 days?
Because OK, I should be able to recover them, especially 2FAuth, but that may happen again when the new update will be deployed.
Sounds like a permissions issue. Make sure the user that runs Homepage has at least read/write access to the host path volume. According to the Run-as Context in the UI, the default UID:GID for Homepage is 1000:1000.
I have no 1000 user or group.
I finally succeed to have an working homepage, by creating the docker with Dockge, with a dataset for which I am the owner (UID:GID = 3000:3000), and not UID:GID precised in the config of the docker.
FYI, I tried creating another instance through Truenas pointing the same dataset, and that cannot work, I still got the same error message.
Ok I was just able to deploy Homepage using a host path. You need to use the Enable ACL option to give permissions to the ID 1000, which is the value shown in the Run As Context on the app details screen.
It doesn’t actually matter whether a matching TrueNAS account exists or not, you’re just granting permissions to the container account with ID 1000. You might also need to check the force ACL option if you’re reusing a dataset with existing data on it.
Thanks, but I was finally able to deploy one without precising any ACL in the config through Dockge.
services:
homepage:
image: ghcr.io/gethomepage/homepage:latest
container_name: homepage
ports:
- 3000:3000
volumes:
- /mnt/mystorage:/app/config # Make sure your local config directory exists
restart: unless-stopped
networks: {}
And that works. It is with a dataset created with generic config, and I am able to add a SMB share and then edit the different yaml files. So for now, I will stay like that .
Good that you finally make it work, but it is quite different than it was before ElectricEel.
I have now to figure out how to do the same magic trick with 2FAuth… but it is far more complicated. Maybe it is based on the same kind of specificity.