Hey team - I’m having some troubles understanding how things are sync’d in pulumi. I’ve created some resources via gitlab-ci (using slightly modified example from the docs, which works as expected!), but then went to try and push an update locally and

pulumi preview

wants to recreate everything. I did have to create a “fake” stack in my docker container as pulumi in gitlab wouldn’t recognize my stack (see https://pulumi-community.slack.com/archives/C84L4E3N1/p1602800146017200) which I assume is part of my problem. Am I missing something here or is there something I am doing wrong?

Hi, I am currently evaluating pulumi especially from testing point of view. I was wondering if pulumi does check the input parameters when running

pulumi preview

. For example in a

VirtualMachine

resource with no

admin_password

attribute and

disable_password_authentication=False

I would expect that it will throw an error, but it did not. Thanks

Hello, I'm working through the getting started docs, and I'm doing the kubernetes + go implementation. I'm on the "Modify the Program" step, and I'm getting an error when I run `pulumi up`:

Diagnostics:
  pulumi:pulumi:Stack (pulumi-quickstart-dev):
    # pulumi-quickstart
    .\main.go:48:12: cannot use v.Template (type "<http://github.com/pulumi/pulumi-kubernetes/sdk/v2/go/kubernetes/core/v1|github.com/pulumi/pulumi-kubernetes/sdk/v2/go/kubernetes/core/v1>".PodTemplateSpec) as type *"<http://github.com/pulumi/pulumi-kubernetes/sdk/v2/go/kubernetes/core/v1|github.com/pulumi/pulumi-kubernetes/sdk/v2/go/kubernetes/core/v1>".PodTemplateSpec in return argument

    error: an unhandled error occurred: program exited with non-zero exit code: 2

My

main.go

is copied and pasted from the program on https://www.pulumi.com/docs/get-started/kubernetes/modify-program/

Hello, anyone know of an example where one creates a Kubernetes cluster (it can be in any provider) and automatically hooks it up with Kubectl?

Hi! A quick question, does Pulumi handle diffs internally for Lambda code references, or is it possible to override that behavior using sourceCodeHash?

Trying to do a targeted refresh of one resource but get the error message "Did you forget to escape $ in your shell?" How should I escape this?

I like micro-stacks, that’s how we are doing it at our project, but now we are facing new issue, how to manage cross-stacks dependencies? As an example, if I’ve updated fqdn at serviceA, how serviceB will know about the update?

I'm looking to build 2 docker images for my project. One that only rebuilds when my dependencies changes and one for my project which uses the first image as a base. Any suggestions on how to achieve that?

Love the tooling and love how supportive everyone is here.

Q: From my understanding Pulumi stacks' secrets aren't being shared with each other, could anyone please point me to the codebase where the KMS-related secrets logic reside? I am hoping to dig a bit into it to see if it is possible to share a same of secrets between many stacks. Many thanks in advance

I'm trying to run

repository.buildAndPushImage(inputs.buildOptions)

in the

create

method of a

pulumi.dynamic.ResourceProvider

but I'm getting the following error:

Error: Program run without the Pulumi engine available; re-run using the pulumi CLI

. Is that not allowed?

I can't seem to wrap my head around service roles and ElasticSearch - please help 🙏 So service role can only be created once, and it's data is stored in pulumi state. However, deploying the same project from a fresh git clone will yell at you if you try to create that same role again. I tried creating conditional logic (see the image below). Now I'm getting this error which doesn't say much. Has anybody done this successfully? Am I doing something wrong here? Thanks a lot! (I saw some issues on Pulumi Github but none of them helped, some are still opened).

I am using pulumi to create a GKE cluster, and kubernetes secrets. Creating the secret fails with

configured Kubernetes cluster is unreachable: unable to load Kubernetes client configuration from kubeconfig file: invalid configuration: no configuration has been provided, try setting KUBERNETES_MASTER environment variable

- how do I get pulumi to use the configuration that it has created for the cluster?

I'm getting the following error when creating a User Managed Identity:

azure-nextgen:managedidentity/v20181130:UserAssignedIdentity (nvp-ppi-avlx):
    error: azure#getURLFromLocationHeader: invalid polling URL '/subscriptions/****/resourcegroups/nvp-ppi-windows-rg/providers/Microsoft.ManagedIdentity/userAssignedIdentities/nvp-testresourcename': StatusCode=0

This is the code I am using:

var userAssignedIdentity = new Pulumi.AzureNextGen.ManagedIdentity.V20181130.UserAssignedIdentity(miName, new Pulumi.AzureNextGen.ManagedIdentity.V20181130.UserAssignedIdentityArgs
{
    Location = "West US",
    ResourceGroupName = "nvp-ppi-windows-rg",
    ResourceName="nvp-testresourcename"
});

I am able to use 

az identity create --resource-group nvp-ppi-windows-rg --name nvpppitest

. My resource group is in 'West US'... Any ideas? Interestingly, the identity does get created - I just get a message from Pulumi that it failed. Also, I will add that testing with the non-AzureNextGen provider (Pulumi.Azure.Authorization.UserAssignedIdentity) adds the identity correctly.

Hi, I’m new to Pulumi but I’m loving it so far. I have a question regarding EKS + Kubernetes + ELB + SSL 🙂 : I’ve created a k8s cluster backed by EKS, I’ve added a web service and I’ve exposed it as type “LoadBalancer”. With this configuration, Pulumi has created an ELB for me and I can successfully reach the service. The problem is that now I want to add an HTTPS listener to the ELB but I don’t have a reference to the ELB in the code. So: • Is it possible to tell Pulumi to create an ELB with http and https listeners when I expose a k8s service as type “LoadBalancer”? • or, is it possible to create the ELB first and then pass a reference to the service so that it uses it instead of creating a new ELB? • or, is it possible to expose the service as “ClusterIP” and then create an ELB in the same vpc of the k8s cluster and make it point to the private IP? • or, is there a better way to do this?

Is there a way to wrap lines in the

pulumi up

terminal output? When I run #kubernetes up commands, I have to zoom all the way out to see the status message. I guess I can go to app.pulumi.com but doesn’t the defeat the purpose of having the terminal output. It’s easier in general to keep an eye on the terminal output for my personal workflow.

pulumi is trying to delete things that don’t exist. A

pulumi refresh

(which presumably would inform it that the thing doesn’t exist) times out. What’s the general approach for handling this situation?

Hi, I'm new to Pulumi and I'm with difficulty to integrate an WebApp (Azure Function) to a VNet. I'm using Azure Next Gen and after creating the Azure Function i'm creating a WebAppSwiftVirtualNetworkConnection, but when I deploy my IaC I get the error "cannot create already existing resource", it looks like it's some conflict with the Azure Function already created. This is the right way to integrate a VNet with a WebApp? I can't find exampes using Azure Next Gen :(

How do I tell pulumi to not try to delete something? I added

protect: true

figuring pulumi would skip over it when I do

pulumi destroy

- but instead that halts the entire destroy process. I just want pulumi to skip over it when destroying the stack

hello everyone, I use pulumi get a default VPC and create a aurora serverless mysql attach with default VPC, the default VPC has 6 public publicSubnets and has no privateSubnets, but when I pulumi up, it create 2 VPC endpoint, how to prevent pulumi to create the VPC endpoint?

I also need to create a ECS farget task to execute a docker to access Internet, so I plan create it in default VPC, so it can access the internet and can access the auroa serverless mysql

Hey! I would like to evaluate Pulumi as an alternative to Helm. I've several deployments running with Helm already, is there any guide about how to this migration and stay alive 😅? My helm charts are saved alongside my code and not stored in any repository

What is the new way of attaching managed policy now that it's deprecated? This is how I've been using them:

/**
     * Attach Lambda Policies
     */
    const lambdaVpcAccess = new aws.iam.RolePolicyAttachment(
      `${stack.prefix}-lambda-vpc-access`,
      {
        role: role.name,
        policyArn: aws.iam.ManagedPolicy.AWSLambdaVPCAccessExecutionRole,
      }
    );

is there a way to run pulumi without any kind of state at all? (no

pulumi login

or

pulumi stack init

) I’m trying to imagine replacing helm with it.

I want to add Log level Errors to API gateway , but I can't seem to find how to pass a role or a managed policy to

awsx.apigateway.API

so when I pass the LogGroup ARN to the

stageArgs

it doesn't generates the following error:

error: 1 error occurred:
        * updating urn:pulumi:dev::booksloth-api::aws:apigateway:x:API$aws:apigateway/stage:Stage::booksloth-api-dev: 1 error occurred:
        * Updating API Gateway Stage failed: BadRequestException: CloudWatch Logs role ARN must be set in account settings to enable logging

This is what I tried so far:

how do I decrypt password I encrypted in config file ? Forgot what it was...

Is there anyway to encourage Pulumi to run DFS rather than BFS through the dependency graph? (e.g. I want to walk a single host at a time rather than all hosts at once)

is there a way to get the URL of a newly created SQS queue and pass it to another resource? Trying to use

aws.sqs.getQueue

, but it accepts a

string

queue name, rather than an

Output<string>

, and therefore runs before the queue is actually created even if i use a static name for the queue.

I have been digging around for a bit and saw this reference in documentation:

Although the template uses a very simple package structure, by placing

__main__.py

in the root directory, Pulumi fully supports properly modularized Python programs and

setup.py

files. This is important if you ever decide to turn your Pulumi program into a library.

But I think that is it. My question: Is there a way to make the

__main__.py

of a pulumi stack a

pip

installable thing? Basically, if I want to build out an application that has heavily tied together processing pipelines and infrastructure, can I tell the

Pulumi.yaml

file "hey please go look for this script that was pip installed" so that I can match up the processing pipeline versions and the infrastructure versions? Or is this a bad idea?