After using Vikunja for a good time (something like a year) by hosting it on a RaspberryPi using Docker, I have decided today to update the images as it was a while.
After that update, I had an issue telling me that the migration of the DB failed (I haven’t noted the error message, sadly). The API image was failing and restarting permanently.
I’ve tried to rollback to older versions, then the migration seemed to work and the applications started properly.
Though, when I’m trying to connect to any account, I now get Wrong username or password.
I’m sure of the identifiers since I’m using a password manager who was still working before the update.
Any idea of what could have gone wrong there?
Thanks in advance!
Anything in the logs? How are you hosting Vikunja? Can you share your config?
The logs mainly contains the actual error.
Besides that, only the first line of log puzzle me a bit but it may be legit:
api_1 | info: creating the new user vikunja with 1000:1000
api_1 | usermod: no changes
api_1 | 2023/10/31 13:28:37 No config file found, using default or config from environment variables.
api_1 | 2023-10-31T13:28:37.708378962Z: INFO ▶ migration/Migrate 052 Ran all migrations successfully.
api_1 | 2023-10-31T13:28:37.708839364Z: INFO ▶ models/RegisterReminderCron 053 Mailer is disabled, not sending reminders per mail
api_1 | 2023-10-31T13:28:37.709532949Z: INFO ▶ models/RegisterOverdueReminderCron 054 Mailer is disabled, not sending overdue per mail
api_1 | 2023-10-31T13:28:37.709999315Z: INFO ▶ cmd/func25 055 Vikunja version v0.21.0
api_1 | ⇨ http server started on [::]:3456
api_1 | 2023-10-31T13:28:57.749061669Z: WEB ▶ 192.168.1.110 GET 404 /info 111.795µs - Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) vikunja-desktop/0.21.0 Chrome/114.0.5735.134 Electron/25.2.0 Safari/537.36
api_1 | 2023-10-31T13:28:57.756502682Z: WEB ▶ 192.168.1.110 GET 200 /api/v1/info 1.025934ms - Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) vikunja-desktop/0.21.0 Chrome/114.0.5735.134 Electron/25.2.0 Safari/537.36
api_1 | 2023-10-31T13:28:57.770896677Z: WEB ▶ 192.168.1.110 GET 200 /api/v1/info 716.53µs - Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) vikunja-desktop/0.21.0 Chrome/114.0.5735.134 Electron/25.2.0 Safari/537.36
api_1 | 2023-10-31T13:29:05.650005686Z: WEB ▶ 192.168.1.110 OPTIONS 204 /api/v1/login 29.87µs - Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) vikunja-desktop/0.21.0 Chrome/114.0.5735.134 Electron/25.2.0 Safari/537.36
api_1 | 2023-10-31T13:29:08.079405225Z: ERROR ▶ v1/Login 0ad Wrong username or password
api_1 | 2023-10-31T13:29:08.080022256Z: WEB ▶ 192.168.1.110 POST 412 /api/v1/login 2.426501922s - Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) vikunja-desktop/0.21.0 Chrome/114.0.5735.134 Electron/25.2.0 Safari/537.36
I suppose it’s the default config, then.
How to host:
I’m just using a docker-compose file with a service for the db, one for the api and one for the frontend.
The only thing I’ve change is the image of the api:
vikunja/api:0.21. (I don’t know what was the previous, I was using latest, but I don’t know exactly from when).
It does not contain any login information (except for the DB, but I don’t think that’s related).
Is any of this helpful?
Can you share your docker compose file?
Sure, here it is:
command: --character-set-server=utf8mb4 --collation-server=utf8mb4_unicode_ci
Not entirely sure but the env variables for the api container might need to be uppercase. Can you check if there is data in the mysl database?
Hmm, I’ve tried to uppercase them and it fails with a SQL migration error.
For what it’s worth, it worked like a charm for the last year with this docker-compose.
Regarding the database, it seems there is some data in it. There are many files, for a total of more than 200MB. Not sure what I should look at or how to read these data without the app, though
Are there any tasks? Check with
select * from tasks;
Yes, they seems to be there
Can you enable db logs and check which queries are executed prior to the wrong password error?