Where is the "current stack" saved (for instance when running

pulumi stack select foo

)? How can I see which stack is currently selected?

pulumi stack ls

should show a

*

next to the current stack

That works. Where is that saved? Seems to be

~/.pulumi/workspaces

, not entirely sure I understand how workspaces work tho. Thanks 🙂

now you’re asking the hard questions 😆 I imagine somewhere in

~/.pulumi

FYI most (if not all) stack related commands in the CLI accept a

--stack

or

-s

argument to specify the stack you want to operate on, if you want to forgo messing with what stack is currently selected

I would have liked to have the equivalent of

terraform workspace show

but

pulumi stack ls

will work for now (albeit a bit slow). I do use the

--stack

command sometimes, especially since all my stacks are in the same project (because of https://github.com/pulumi/pulumi/issues/2814)

That issue doesn't force all your stacks to be in the same project, it just means that project name is not part of stack identity with S3 backend. You could have separate projects if you wanted, you would currently just need to namespace your stack names by project name

I guess I confused the word project, they are in different projects (i.e. different folders with a Pulumi.yaml each), but

pulumi stack ls

still returns all the stacks of all the projects that share the backend. So I could technically have

.
├── a
│   └── Pulumi.yaml
└── b
    └── Pulumi.yaml

and go into

b

and do

pulumi stack select a.dev

ah yea, I guess in that sense they are all in the same project in that all stacks are always accessible. my bad

No worries! I was kinda hoping the current selected stack was easily accessible (say an environment variable) so that way I could at least have a different TTY in each folder, each with a different stack selected. Right now I either have to use

-s

on every command, or change selected stack in each folder (but that's very error prone)

when you're developing locally it's definitely a pain. But I think when you're setting up your pipelines and automation just specifying

-s

on everything is a good solution