Repository navigation
New deploy: handle deployments to other target projects than is currently checked out #1069
Description
Not sure if this is real yet, taking notes
- I'm working from staging and local (
3756-propagate-project-credentials) - I'm using new CLI pull, but I don't think that's material because all the provisioner API stuff is shared with the prod deploy package
- I have successfully pulled a sandbox from staging and expanded it locally (cool!)
- When trying to deploy the same to local, I get a 403 error
- The project is set up locally, and I can pull it perfectly fine
- I can't see any error in lightning
Activity
josephjclark commentedon Oct 15, 2025
When this stuff happens the errors we get out of lightning AND the CLI are just so poor. It HAS to be a priority to give really good feedback when things go wrong.
I am a bit concerned that we don't know or say the difference between "this project genuinely doesn't exist" versus "this project exists but you don't have access to it"
josephjclark commentedon Oct 15, 2025
Ok, I think the problem lives in CLI deploy, and because the new project stuff isn't using the right id
josephjclark commentedon Oct 15, 2025
Yes, Ok, I see the problem. This might be a new problem that needs thinking about.
project@app.staging.openfn.org.yamlopenfn deploy --beta <local-uuid>I can't just pull from one server and deploy to another, the UUIDs are all wrong. This is actually a merge.
I think what really needs to happen here is:
deploy projectid, CLI needs to check and see if that project exists alreadyOf course version tracking needs adding onto all this too.
So basically - the new CLI deploy is barely done at all, and doesn't take merging into account at all.
[-]Unable to deploy?[/-][+]New deploy: handle deployments to other target projects than is currently checked out[/+]