Error after upgrading to 0.18.1 & 0.18.2

[retiolus]

After upgrading to latest 0.18.1 Lemmy Yunohost version, my Lemmy server look like this:

[retiolus]

The issue is coming from the lemmy service, I can't get logs, yunohost-api crash when trying to get them.

[mdeabreu]

Do you still have this issue? I just tried upgrading on my live instance and noticed the same thing however I was able to dig a little deeper. I think this is if you have a large database. If you can ssh into your machine (or have physical access), you can see what the lemmy service is doing by looking at the /var/log/lemmy/lemmy.log file. In my case I saw the following line:

lemmy_db_schema::utils: Running Database migrations (this may take a long time)...

Yunohost thinks the upgrade succeeded but you should have seen a warning saying the lemmy service didn't fully start, that's because it's still running the migrations.

You can follow along by running tail -f /var/log/lemmy/lemmy.log and waiting until it says something new. You can also run top and see that postgres is using a ton of CPU as it goes through the migrations. It seems like the best bet is to wait until the migrations are complete and you see Database migrations complete. in that log file.

[retiolus]

That's correct

[retiolus]

2 h and still doing the database migration (~7 GB database), is this normal?

[retiolus]

Database migration finished and still having the "Error" message.

API error: FetchError: request to http://0.0.0.0:8536/api/v3/site? failed, reason: connect ECONNREFUSED 0.0.0.0:8536

[da-am]

I'm seeing this too. After migration I'm getting:
Error connecting to postgres://lemmy:9WVl5bVF0FA9myk1slnMFlY7g9KEsW@localhost:5432/lemmy: could not connect to server: Connection refused

[mdeabreu]

I was seeing the connection refused to the postgres server, attempting to start lemmy again got me past that issue. But then it appeared that I was having troubles federating. I wasn't able to see any incoming activity.

I've opted to restore from my backup for the time being.

[retiolus]

Fixed the issue following these steps:

• run the following SQL query to remove server icon: Server error after updating to 0.18.0 LemmyNet/lemmy-ui#1530 (comment)
• update environment variables in /etc/systemd/system/lemmy-ui.service:
  • LEMMY_INTERNAL_HOST --> LEMMY_UI_LEMMY_INTERNAL_HOST
  • LEMMY_EXTERNAL_HOST --> LEMMY_UI_LEMMY_EXTERNAL_HOST
  • LEMMY_HTTPS --> LEMMY_UI_HTTPS
• restart lemmy-ui service

Updated config of the environment variables: https://github.com/LemmyNet/lemmy-ui#configuration

[jahway603]

Thank you @retiolus for this possible fix, as I do not see the OP error. It's not clear if the server is now federating or not though.

[jahway603]

Now I'm getting the original error again. Anyone have any ideas on how to fix this?

[retiolus]

My server running 0.18.2~ynh3 was completely broken, here is how I fixed it (not sure if this will allow futures updates from YunoHost):

node: v20.14.0
yarn: 1.22.22

1. sudo -u lemmy bash
2. cd
3. delete lemmy-ui: rm -rf lemmy-ui
4. git clone https://github.com/LemmyNet/lemmy-ui.git --recursive --> --recursive is very important
5. cd lemmy-ui
6. git checkout 0.18.2
7. yarn install --pure-lockfile
8. yarn build:prod
9. here I don't know why but my lemmy server was running on port 8537 (default is 8536), even changing lemmy-ui environment variables I didn't manage to change its config so...
10. sudo nano /var/www/lemmy/config/config.hjson and change port to 8536
11. sudo systemctl restart lemmy
12. sudo nano /etc/nginx/conf.d/lemmy.cat.d/lemmy.conf and change all 8537 to 8536
13. sudo systemctl daemon-reload && sudo systemctl restart nginx
14. sudo systemctl restart lemmy-ui