did you copy the new internal connection string from postgres & redeploy?

(i believe it would create a new one)

I've checked and made sure that DATABASEURL in my compute service is the same as DATABASEURL in the postgres service.

I want to click on Somethings wrong, but doing that will just revert the migration:

I tried to re-deploy the current instance(which is btw connected to the database), but it fails:

I am able to connect to the db using the DATABASEPUBLICURL locally though so the db is actually running.

I have had zero problems with this until I decided to change region

im not really sure if this would work, as i havent done a swap from region to region db wise, but you can deploy a new db in the correct region, and then reconnect it to your backend service (and then deattach the volume from the original one, and readd it to the new one, ensuring all of its vars are in the service)

Ive tried that, I remember when I used railway several years ago there was a way to detach the volume but I cant find it now:

hmm that im not sure why it doesnt appear

maybe because im on free plan?

perhaps it might be bc you tried to swap the region already

i dont think so

I could do the pro plan but i kinda wanna be sure things works then 😄

nah i dont think its a plan issue

Tried creating a completely fresh pg db and still no such setting:

hmm..I wonder then if it is 🤔

I assume you right clicked it the actual volume, not the db?

Yes I right clicked the volume.

technically you can delete the db and it’ll unmount the volume but that’s not really a fix

but deleting the db means you’d be losing your data in it

When I deleted the db, the volume is now a separate block.

so Ill try attaching it to a new postgres service

omg, but now the new db i created ends up in singapore again

within your account settings, is your default region Singapore or Asia?

aha, thx

try that and deploy a new db and it should work with your region

Still failed

I even tried now to re-create the compute service and it's still fd up xD

oh it must be the freakin volume that is some how corrupted.

I made a completely new db now and connected it to the completely new compute service and now it correctly works.