The same AWS credential works on my other pulumi projects.

Can anyone help me debug this?

I've got a situation I've not seen before; I'm trying to install linkerd with pulumi with the helm chart -- I'm passing in the certificates as secrets from the pulumi config and interpolating them in the

values

section of the

k8s.helm.v2.chart

resource The issue is that every-time I run a pulumi up it looks like the secret values are changing in k8s and the checksums on the deployments are changing; any thoughts as to why?

Hi, I'm running into an issue using the Pulumi Github action and I'm not sure what is going wrong. Could somebody help me make sense of this? 🙏

[resource plugin gcp-5.0.0] installing
  
  Downloading plugin: 0 B / 34.41 MiB    0.00%
  Downloading plugin: 3.51 MiB / 34.41 MiB   10.19% 1s
  Downloading plugin: 10.03 MiB / 34.41 MiB   29.16%
  Downloading plugin: 15.23 MiB / 34.41 MiB   44.25%
  Downloading plugin: 20.62 MiB / 34.41 MiB   59.91%
  Downloading plugin: 25.93 MiB / 34.41 MiB   75.36%
  Downloading plugin: 31.06 MiB / 34.41 MiB   90.26%
  Downloading plugin: 34.41 MiB / 34.41 MiB  100.00% 1s
  
      pulumi:pulumi:Stack replayvalue-dev running 
      pulumi:providers:gcp default_5_2_0  error: no resource plugin 'gcp-v5.2.0' found in the workspace or on your $PATH, install the plugin using `pulumi plugin install resource gcp v5.2.0`
      pulumi:pulumi:Stack replayvalue-dev  
      pulumi:providers:gcp default_5_2_0  1 error
   
  Diagnostics:
    pulumi:providers:gcp (default_5_2_0):
      error: no resource plugin 'gcp-v5.2.0' found in the workspace or on your $PATH, install the plugin using `pulumi plugin install resource gcp v5.2.0`

Hello everyone, happy Friday… something happen on one of my

dev

stacks and become

dev*                                  n/a           n/a

I’m trying to import the backup from

dev.json.bak

but getting

error: could not deserialize deployment: unexpected end of JSON input

from `pulumi stack import --file dev-bkp.json`… is there a way to directly import the

bak

or something else beforehand should be done?… also

pulumi refresh

doesn’t do anything…

Getting

pulumi stack output

in a file using the Azure DevOps Pipeline task We need to do some post processing using the stack output on azure devops. I want to do this with a service connection to avoid having to worry about secrets. If I run pulumi using the azure CLI task I'm getting an error relating to connecting to the blob container that's used for the pulumi back end. This is notionally easy to fix, run

pulumi stack output --json

using the pulumi task (which I know will work because

pulumi up

works) capturing the output to a file that I can then feed into the powershell that does the things we can't currently do with pulumi. Except I can't capture the output 😢 Open to suggestions at this point (before I put a feature request in to add --output to the CLI)

Anyone knows if there is a known issue showing diffs properly in certain scenarios? Been trying to search for open issues but no success so far. I first thought it would be output/input related but it’s occurring also for values that are just plain strings 🤔 (using typescript, pulumi v2.24.1)

Hi, sorry for posting the same issue again. Now I have more details. Two different projects, on the same machine, almost the same Pulumi yaml file. It was successful on the left side but failed on the right side because of some account issue.

The problem AWS account

arn:aws:sts::228886564947:assumed-role/AmazonLightsailInstanceRole/i-04f5fe371a036054d

seems to be an account that running the light sail, It's even not my account. I am running the deployment on a light sail instance. My understanding is instance role is of lower priority than the configuration in the

.aws

folder. What makes the project on the right side to choose the instance role?

Thank you Tenwit, please see my new post in the channel

Hi all - I'm still new to Pulumi, and probably getting over my head a bit. But, I'm looking for a way to use Pulumi to drive my setup of cloud based software such as FiveTran. My approach was to extend ResourceProvider and implement the create method, calling a REST API using the python "requests" library. However, when I do this, I get an exception:

RuntimeWarning: coroutine 'invoke.<locals>.do_rpc' was never awaited

Here's the question on stackoverflow, in case anyone has an answer or suggestion as to what I'm doing wrong here: https://stackoverflow.com/questions/67335936/how-do-you-write-a-pulumi-provider-that-makes-a-rest-api-request-in-pytyon

Hey folks, I am interested in evaluating Pulumi for my organization, but have a question surrounding provider credentials. In my ideal workflow I would like to provision a very capable IAM role for Pulumi but prevent those credentials from being used/abused outside of the Pulumi plan/apply workflow. Does anyone know if there is a simple way to provide a remote-run/remote-execute environment for Pulumi? Or any way to shield/hide these powerful provider credentials? Thanks in advance! 🙏

Hi guys, wondering how to get existing ALB in Pulumi

<https://www.pulumi.com/docs/guides/crosswalk/aws/elb/>

# what is the equivalent code for existing ALB?
const alb = new awsx.lb.ApplicationLoadBalancer("web-traffic");

Is there a way to enable

ephemeral containers

in a kubernetes cluster created using eks?

hello

Just wanted to check , when pulumi will start supporting Java ?

hello. i'm not getting invitation emails, email verification mails from pulumi. when i click on the resend verification email button on my email address in pulumi, i never receive any email. how do i fix this?

I have a lot of

dangling

Pulumi resources in my (TypeScript) stacks, resources which are created but not used as input to any other resource and which are neither added to any output nor returned from the main inline function (using the automation api). Looking at many Pulumi examples this seems to be a common pattern. Think of something like this:

function createBucket() {
   // Create AWS S3 Bucket but do neither output nor return it. It is also not used 
   // as input to any other resource.
   new Bucket("myBucket", {...});
}

However, rather often I observe that such resources are not created on first stack deployment resp. are delete on later stack updates. It looks like it depends on some “race condition” (timing issue) whether those resources are registered by the Pulumi runtime or not. Usually adding such

dangling

resources as dependency to another resource (using

dependsOn

), or alternatively making it a stack output resp. returning it from the main inline function resolves that (racy) problem. Now I am wondering, is this expected behaviour when using Pulumi with TypeScript/NodeJs or is this a bug in Pulumi?

Is there a tool already that given a state file will generate the code for it?

Hey, how do i use an existing ecs cluster by cluster id?

My heretofore fully-working project is now complaining that Error: The root stack resource was referenced before it was initialized. What sort of thing should I be looking for in diagnosing this?

when using integration tests, how do I decode stack outputs that are secrets?

nm I have to upgrade my testing library

Who can help me setting Azure pipeline with self-hosted Pulumi state in Azure blob ?

Hi, let's say I have a part of my infrastructure provisioned by Pulumi while the rest is not. Some services from the non-provisioned part need to access resources that are Pulumi-managed (e.g. they need to call Lambdas). What's the best / most aligned way how to reference those provisioned resources? What comes to my mind: • Name those Pulumi-provisioned resources explicitly (i.e. don't use auto-naming). • Use auto-naming but then store the names somewhere (e.g. SSM Parameter Store, but then I'd need to name the parameters' paths explicitly). • Maybe just stick to knowing which project/stack the resources are managed by and somehow ask for the outputs (but I don't know how to achieve this at the moment). Thanks for hints!

Hello there - I am using aws-typescript. Doing first basic provisioning of EC2 instances, and the name of the instances is not getting set on creation? I would assume the resource name is used as the VM name? no other args to tap into to be able to set the EC2 instances name either as far as I can tell.

Hi, had one question about AWS ECS. According to this article, it is possible on ECS to run a standalone task. But it seems that pulumi only supports attaching task to a service. Does anybody know if Pulumi can do that ?

Is yarn 2 supported?

Hey all, is there a Pulumi module available somewhere for setting up all the required GCS resources to host your own state?

👋 hello! qq: what’s the recommended approach of reading a secret from another project? context: my project

infra

exports a

kubeconfig

for a kubernetes cluster (and it’s a secret), and my project

services

reads it, creates a

k8s.Provider

and uses that to provision my services in the cluster. I cannot use

StackReference.getOutputValue()

because the output is a secret. If I use

StackReference.getOutput()

, then I have to do

Output.apply()

to transform it into a

Output<k8s.Provider>

. But the k8s resources only takes `k8s.Provider`s, not `Output<k8s.Provider>`s… so.. I guess one option is to put everything inside the

apply()

call.. but that’s undesirable. I’m wondering if there’s a cleaner way