Hello,
I spun up my instance and when I was doing the initial setup I got an error "Server already exists". I reloaded the page and realized that the server was a copy of https://exploding-heads.com/. I was really confused since I didn't do anything. I didn't even federated with that instance. After that I changed the settings to how I wanted them to be and moved on with the day. I thought it was one time issue.
Few days later I restarted the server and when I got back to it the name, logo, description and sidebar have changed back to the ones from the other instance. I changed them back to normal and decided to restart just the docker containers. Again the same issue.
Normally the logs help but this time nothing out of normal shows up.
We had a bug like this in a previous Lemmy version, but that was fixed. The thing is that name, description etc for the local instance and for federated instances are all stored in the same db table. To read info about the local instance, Lemmy reads the row in
sitedb table with lowest id.So it would be helpful if you could post the content of that table, something like
select id, name from site order by id.Hmm, this is really strange. I decided to try to restart both the containers and the server several times. everything seems to be working now. Also I don't know how to access the table.
sudo docker-compose exec postgres psql -U lemmy
Anyway its strange that a restart would make any difference. When we had this problem, the site name etc changed with every page reload (as it was reading a random row from the db table each time).
After this command what do I do?
Run the sql query from above:
select id, name from site order by idI ran it already. Nothing happens
Ah you need to add a ; at the end.
Looks normal then, cant think of anything that would cause your problem. I guess when it happens again, you can run the same sql query, and check that your instance name still has id 1. But it would be very strange if that changed somehow.
Ok, thank you really much! I will report if something goes wrong again.
No problem. This PR would fix the problem once it gets implemented.
https://github.com/LemmyNet/lemmy/issues/2365