Hey! Railway had an incident earlier today, can you try redeploying the affected services? DB and N8N

I have tried earlier several times.

Sync is still failing - dismounts existing volumes in the target location and fails the re-deploy with no volumes creaated.

(Sync from staging -> production)

As for staging env, tried to redeploy now and seems to work fine, can access n8n dashboard again.

My suggestion was just to get your services up and running, are you still having the sync issue? I'll escalate this thread if so

What error does it throw on fail?

Yes, the sync issue, still persists.

The sync has the addition of volumes listed in the deployment changes.

But then the volumes are not created nor added and I'm left with message in "build logs" - This service requires a volume to be mounted at /var/lib/postgresql/data. Please attach a volume to the service and try again.

This happens regardless of whether the services have volumes connected or not. If there are connected volumes, they get disconnected.

However, I cannot find any error, that would be directly relating to this.

The image shows the changes, which are supposed to be made.

Assuming I start from a blank project, how do I reproduce this?

To be honest, not entirely sure.

I have tried to delete everything and wanted to sync staging -> production, but I didn't find the "Sync" button available anywhere.

So, I have tried with minimun, which was:

1. Create environment.
2. Deploy empty service.
3. Sync environments.

That would lead to the changes copying services from "staging" and removing the empty service.

However, some services with volumes were deployed successfully, but some were not.

So far I have had this issue only with the n8n services.

Other services - NocoBase, Supabase have successfully synced and volumes appeared.

But the volumes don't seem to have synced properly.

I have tried to access the synced Supabase deployment, but the dashboard doesn't load - error "This project doesn't exist.", which points at the fact, that the synced Postgres and its volume don't contain the data from "staging" deployment.

The sync button is not available in the empty environment, so I had to do it with workaround like this, but empty service shouldn't really be affecting the result.

I would love to help here, but due to volume, we would need a fully reproducible example in order for us to spend some cycles on this.

I will try to remake and sync the environment again later.

Which in itself is not an issue, because services can be synced individually via your "Sync from another environment" function accessible via right-click on the service under "Config" dropdown.

I have tested this also aside the environment-wide sync.

But same issue - the volume is not being duplicated/synced.

And volume itself doesn't have the option to be synced or copied over to the other environment.

Yes, that is correct understanding.

I have used the term "copy" here as description of what I expect the Sync function to do.

Which in this regard means to duplicate (copy/sync - your pick for the wording) all services and data from existing "source" (the one we sync from) environment or service, as is, into the "target" (the one we sync into) environment/service.

This includes the connected volumes and the data they contain.

As per screenshots - the functionality is actually trying to do that, but with zero result.

The only exceptions are public domains or other elements, which would cause direct conflict if same values were applied across multiple services/environments.

Oh I see what you mean.

We do not support copying the contents of the volume over, and that's unfortunately not something we will have in the immediate future.

Then the question would be, why is the sync trying to do so or puts it into deploy changes to begin with?

That's kind of misleading.

On the other hand, if the contents of the volume cannot be copied over like this - is there an alternative way to do so?

The sync is creating a volume. I don't think there is any misrepresentation that it will copy data.

You would have to implement something on your end to copy over the data.

Well, then there is still issue, because the volumes are not getting created reliably. 🤔

Some of the services got the volumes, but some didn't even after several tries.

Then we are back to this unfortunately.