May ... I've not had a build green since May 😞

I'm trying to use a secret as the kubeconfig for the k8s.Provider, but it keeps failing. Does anyone know what I am doing wrong? 😞

const kubernetesProvider = new k8s.Provider("metrics", {
  kubeconfig: config.requireSecret("kubeconfig")
});

I'm now doing

pulumi config get kubeconfig > kubeconfig

and loaded this file during

pulumi up

. Please tell me there's a better way 🙂

OK. Found some sort of problem 😕

const kconfig = fs.readFileSync("../kubeconfig").toString();
const kubernetesProvider = new k8s.Provider("metrics", {
  kubeconfig: kconfig
});

Fails:

error: unable to load Kubernetes client configuration from kubeconfig file: invalid configuration: no configuration has been provide

But

kubectl get pods -n community --kubeconfig=./kubeconfig

works 😞

Trying to create a lambda for handling S3 object creation:

transcriptionJobs.bucket.onObjectCreated('publish', async ev => {
  const s3Client = new aws.sdk.S3()
  const kinesisCliet = new aws.sdk.Kinesis()
  for (const { s3: { bucket, object } } of ev.Records || []) {
    console.log({ bucket, object })
  }
})

and getting an error:

Error: failed to register new resource publish [aws:lambda/function:Function]: 2 UNKNOWN: failed to compute asset hash: asset path 'node_modules/@lemonade/infra-common' is a directory; try using an archive

Arch AUR updated to 1.4.0 🙂 https://aur.archlinux.org/packages/pulumi-bin/

Following this guide: https://www.pulumi.com/docs/guides/crosswalk/aws/ecs/#creating-an-auto-scaling-group-for-ecs-cluster-instances I am calling

createAutoScalingGroup

on an

awsx.ecs.Cluster

. I get the error

Property AvailabilityZones cannot be empty

. I don’t see an option for setting

availabilityZones

through

createAutoScalingGroup

.

Is there an easier way to copy a stack config with secrets? I am creating a new stack and right now I just copy the yaml and redo all the secret values with

pulumi config set --secret ...

what is the Windows PS equivalent to the instructions at https://www.pulumi.com/docs/get-started/aws/review-project/

Does pulumi do anything to change the output so that CI systems like CodeFresh don’t kill the build due to inactivity?

Is there any way one could re-deploy an existing ECS service with pulumi, like

aws ecs update-service

? The image is being built by CI not using pulumi

Ok this should be easy but how in the world do I get a value out of the Configuration values?

is there any example of codebuild project for githhub private repos. The example works fine when using public repos. I did configure oauth however i get timeout when building code in private repos. IF i configure the same job from ui it all seems to work fine.

Can anyone confirm if pulumi actually works in Windows? Here is what I tried to get started: choco install pulumi (OR upgrade) pulumi version (result: v1.4.0) python3 https://console.aws.amazon.com/iam/ created user, downloaded csv Download the AWS CLI MSI installer for Windows (64-bit) aws --version (result: aws-cli/1.16.266 Python/3.6.0 Windows/10 botocore/1.13.2) python --version (Python 3.8.0) pip3 --version (result: pip 19.3.1) aws configure (key from csv) mkdir quickstart (pulumi-qs) cd quickstart pulumi new aws-python pip install virtualenv virtualenv -p python3 venv virtualenv venv . .\venv\Scripts\activate pip3 install -r requirements.txt Result: Building wheel for pulumi-aws (setup.py) ... done Created wheel for pulumi-aws: filename=pulumi_aws-1.7.0-cp38-none-any.whl size=1544097 sha256=xxx Stored in directory: C:\Users\gunna\AppData\Local\pip\Cache\wheels\e7\18\87\id Building wheel for grpcio (setup.py) ... error ERROR: Command errored out with exit status 1: command: 'c:\users\gunna\pulumi-qs\venv\scripts\python.exe' -u -c 'import sys, setuptools, tokenize; sys.argv[0] = '"'"'C:\\Users\\gunna\\AppData\\Local\\Temp\\pip-install-mf51566e\\grpcio\\setup.py'"'"'; file='"'"'C:\\Users\\gunna\\AppData\\Local\\Temp\\pip-install-mf51566e\\grpcio\\setup.py'"'"';f=getattr(tokenize, '"'"'open'"'"', open)(file);code=f.read().replace('"'"'\r\n'"'"', '"'"'\n'"'"');f.close();exec(compile(code, file, '"'"'exec'"'"'))' bdist_wheel -d 'C:\Users\gunna\AppData\Local\Temp\pip-wheel-4x1bc1ec' --python-tag cp38 cwd: C:\Users\gunna\AppData\Local\Temp\pip-install-mf51566e\grpcio\

How would I call an async method that I need to export as a value in my stack?

Would it be really difficult to add an attribute to an export so it doesn't show up in output?

I feel like the Docker container used for pulumi/actions is way too large? Why is it over 500MB? I'm gonna see if I can fork that and get it running on Alpine or something...

Another question, is there a way to specify a wildcard in ci.json? For instance if I want to deploy by using Releases in GitHub it ends up throwing an exception No stack configured for branch "refs/tags/v1.0.0": "<stack-name>", but in reality this could just map to master branch...

HI, I just started with Pulumi today with Azure. However, seems .net Core is not supported. Please share it any workaround?

Any chance to have https://github.com/pulumi/pulumi-kubernetes/pull/666 merged? Maybe we can split it into two and at least merge the extended functionality of

transformations

? it's a blocker for using helm charts currently...

[RESOLVED] Using

pulumi.kubernetes.ConfigFile

and disabling kubectl validation for CRD unknown fields (relating to

cert-manager

)

Is there a generic

k8s

resource I can use to create an object of a CRD without implementing a full

pulumi custom resource

? Looking into specifying some

cert-manager

`Issuer`s and `Certificate`s in code without just dropping everything in a yaml file and using

ConfigFile

for it.

Hey all, LOVING the project, been porting my infra over to it as fast as I can. Question (I don’t write much JS/TS ordinarily): is there a cleaner way to make variables from the main

index.ts

available for testing? I’m currently exporting them as needed and importing in my tests, but obviously that really clutters up each stack’s output variables.

Hey guys, I am using Gitlab right now and my Runner is not able to assume a role that I pass it. Instead it uses the system-node role.

I was wondering if any of you had this problem and how you fixed it?

When I run pulumi up from a pod within an eks cluster the pod uses the role that is associated with the worker node (EC2 instance) to deploy applications. How do I go about setting up a service account for pulumi execution. My runner has a service account but I am not sure how to run pulumi up with the service account.

Is the Go test harness supposed to fail on different resource types that have the same URN?

panic: fatal: An assertion has failed: Two resources ('urn:pulumi:fuckit::XXXXX::kubernetes:apps/v1:Deployment::squid' and 'urn:pulumi:fuckit::XXXXX::kubernetes:core/v1:Service::squid') aliased to the same: 'urn:pulumi:fuckit::XXXXX::kubernetes:apps/v1:Deployment::squid'

Is there anyway to move a stack to a different project?