Hi, looks like my stack is behaving weird and causing me grief for every “pulumi up” command in run with following error message:

Do you want to perform this update? yes

Updating (dev):

…………………………………………………………………….

pulumi:pulumi:StackReference (grst):

error: post-step event returned an error: failed to save snapshot: .pulumi/stacks/dev.json: snapshot integrity failure; it was already written, but is invalid (backup available at .pulumi/stacks/dev.json.bak): child resource urn:pulumi:dev::xyz::pulumi:pulumi:StackReference::grst refers to missing parent urn:pulumi:dev::xyz::pulumi:pulumi:Stack::xyz-dev

Not sure why it keep saying child resource is missing parent, any suggestion please?

I’m wondering, are there any way I could create an IAM key, based on the resources that I’m deploying (with only the permissions that i required to deploy those resources).

Hi all. I have been using some Pulumi code to install Zalando Postgres Operator and it worked just fine so far. I installed it OK in a new cluster now, but when I run

pulumi update

again it tries to create again a CRD (operatorconfigurations.acid.zalan.do) that it has already created the first time. I can't figure out how to prevent that. You can see the code at https://ybin.me/p/2162e08ece72e646#iOV2fVY34JpU2d1mAqpxc00ru4SCfGwRdcquitOKupY= - any idea of what I can do to fix? Thanks!

Hi, is there a way to get the contents of Pulumi.yaml during runtime? Without bruteforcing and having to read the file myself, I mean

I my case, using Python, I want to know where the virtualenv is, so I can shell out to a command I need to run

Hi All!

Has anyone run apollo-server-lambda with pulumi ?

I have state in an s3 bucket for a stack I havent touched in a while. When I log into that bucket and do a

pulumi stack ls

it doesn't see anything. I can see files in the .pulumi/stacks/ folder. Did the format change somewhere along the line?

This is a weird one. I've cancelled an update in progress (no resources were to be changed, btw) and now my stack seems to be gone.

error: no stack selected; please use `pulumi stack select` or `pulumi stack init` to choose one

pulumi stack ls

shows nothing.

My backend is S3 and I noticed that my credentials have just expired, might be related. Still no stacks available after renewing, though.

My stack's json file was gone from S3. Thankfully Pulumi creates a bunch of backups and I had versioning enabled, but it's very weird why it was just gone. Ctrl-C looks like a real killer here

Hey guys, nice to meet y'all

I'm starting off using Pulumi on GCP

I had some observations I thought could be useful for the direction of the product

I come from AWS + CloudFormation

One thing I noticed off the bat is how Pulumi uses OAuth natively in order to authenticate to an orchestration server

I wrap my development runtime using Docker and I prefer credential files that I can mount into a Docker volume

Pulumi makes me pass in an auth token w/o conventional

~/.pulumi/credentials

or

~/.pulumi/config

like

~/.aws

has

or click on a link to a website

Also not sure how much I like the

venv/

configuration, since I ran my Docker container as a default root user to template out a PoC the file / directory permissions are different b/w host + container and it's a snag to work that out

I'm also not sure whether Pulumi can pass in env variables and output variables from the stack, I think it can I just haven't discovered it yet, but CF makes this trivial since there's a top-level keyword in the docs

Not sure how much I like the

curl -fsSL <https://get.pulumi.com> | sh

since I wanted to lock down the Pulumi version, but I think there's perf differences between curling the tarball directory and having the script do it for you

Curling the tarball URL directly was too slow to be practical imho

I really wish Pulumi would support Cloud CDN, I think this is a GCP-specific issue as GCP Deployment Manager also raises no backends with this explicitly, vs. AWS's CloudFront CloudFormation resource

Instead React.js deployment docs reference GCP App Engine which at minimum costs $18 / mo. (imho I think that's a ripoff given I'm running a blog using an S3 bucket + CloudFront for a few cents every month)

Hope it helps, I'm sure Pulumi has great things ahead and I'm looking forward to using it

hey @thankful-holiday-28161, would be great if you could thread all this awesome feedback, would make it easier to address and reduces the amount of scrollback in the channel. Some things I can provide as specific points: - the

pulumi login

command will generate a

~/.pulumi/credentials.json

, there's an example of it here: https://gist.github.com/jaxxstorm/b7672567bf76bcf64c41d8150b6afaea - unless I'm misunderstanding, you can grab env variable directly using the standard python tools, we also have configuration: https://www.pulumi.com/docs/intro/concepts/config/ and stack references: https://www.pulumi.com/docs/intro/concepts/organizing-stacks-projects/#inter-stack-dependencies which help with configuring user specific config - I'd love to see some numbers for your download speed/perf differences. the

curl | sh

installation method isn't my personal favourite here. We have docker containers ready to use if you need them: https://hub.docker.com/u/pulumi - not sure what your use case is, but it seems like you're doing static assets with a bucke. There's no specific endpoint for Cloud CDN as you say, but if you add a standard gcp bucket it gets cloud cdn and there's also a backend bucket type which might meet your needs: https://www.pulumi.com/docs/reference/pkg/nodejs/pulumi/gcp/compute/#BackendBucket - you can specify your own venv if you like: https://www.pulumi.com/docs/intro/languages/python/#virtual-environments hope that helps, welcome to the community!

Hi All, on my local

gcloud config list

shows my config, however in

@pulumi/gcp.config

all values are undefined. Any ideas?

When is pulumi going to upgrade to v3 officially? https://www.hashicorp.com/blog/announcing-v3-0-of-the-terraform-aws-provider/