Estebiu@lemmy.dbzer0.com 3 days ago Tried removing the tile-server directory, but it still says it’s unheathly. I’ll try debugging more tomorrow. Gosh dang it, I knew updating at night wasn’t a good idea
Estebiu@lemmy.dbzer0.com 3 days ago Estebiu@lemmy.dbzer0.com 3 days ago Tried removing the tile-server directory, but it still says it’s unheathly. I’ll try debugging more tomorrow. Gosh dang it, I knew updating at night wasn’t a good idea
danielgraf@discuss.tchncs.de 3 days ago
Hi Estebiu, i am sorry for the trouble. Lets try to find out what the problem is. The tile-cache is a simple nginx container. I never experienced any problem with it.
I upgraded two weeks ago. I don’t exactly remember when. No mounted directories. Now that I think about it, the map wasn’t showing up before either. I though it was just my mobile connection that was bad. Yes, I do upgrade my docker containers from my phone when traffic is gridlocked. Anyway, doing a ‘docker system prune a’ when the containers are down, and then putting them back up, doesn’t fix things either…
danielgraf@discuss.tchncs.de 3 days ago
Ok, we need to find out what the culprit is. First we check the status of the tile-cache if there is any:
sudo docker inspect $(sudo docker compose ps -q tile-cache) --format=‘{{.State.ExitCode}} - Status: {{.State.Status}} | Error: {{.State.Error}}’This shoudl tell us if the config is wrong. It should not be, but you never know. If that returns 0: Status: unhealthy then the healthcheck is propably the problem. It calls this command inside the container:
curl -f -s -o /dev/null -H “X-Reitti-Upstream-Url: https://tiles.dedicatedcode.com/” http://localhost/custom/planet/latest/0/0/0.pbf || exit 1So, basicly it tries to fetch a tile from tiles.dedicatedcode.com under this url: tiles.dedicatedcode.com/planet/latest/0/0/0.pbf
Hello again! I managed to fix it. Wasn’t all that difficult tbh, but kind of a pain, without knowing why it broke. Turns out, there’s been a change in the docker compose, and the healthcheck i had in mine wasn’t necessary anymore. I think it was checking an old location where the modern tile-server wasn’t putting anything anymore, so the healthtracker though the tileserver was down, even if it wasn’t. Removed it from the docker compose, and voilà. Thanks