And one more thread on another topic: outputs synchronisation between stacks. In my previous thread I described a situation where my kubeconfig output had some invalid data in the generated yaml kubeconfig (

[object ..]

instead of the actual project id string). This was “consumed” in my another stack via a stack reference and

getOutput

. The issue was that after I have fixed the kubeconfig output in my first stack, the one that was consuming it hasn’t detect any changes and I couldn’t fix the issue reported by k8s provider due to invalid yaml config. Is there a document describing when outputs are being updated in the consuming stacks?

Is there any way to get the current

gcloud

account used by pulumi ? Either the one from

gcloud auth

or the one from

export GOOGLE_APPLICATION_CREDENTIALS

(which drives me crazy to have both btw) ?

gcp.config.credentials

returns

undefined

;

gcp.serviceAccount.getAccount()

or

gcp.serviceAccount.Account.get()

wants an id while I only can know the email of the account.. I’m kind of out of options from what I can think of.

i'm working on an orchestration container that has pulumi, az-cli, aws-cli, terraform, bunch of binaries in a controlled environment with authentication through env vars etc.. anyways, i found that if you

npm install

inside docker, and then on my host macOS, it complains about the version of gRCP.. however, while developing, I found that if you

pulumi up

with the wrong one in place, apparently it deletes everything. Since this is for automation, I added a

-y

to my

pulumi up

command, but what i got was a full destroy. Probably not a bug, per se, but thinking you guys should be aware.

what’s you all’s recommended way to structure the api layer so it works well with local testing and such? In the past, we’ve scaffold out an express app that could run locally and run all the functions. (this was before SAM, mind you). but then deployment…we wouldn’t use express. I’ve used localstack for a lot of local simulation of other parts of AWS too. For Pulumi endpoints … what do you all consider a best practice?

Hello guys, I'm trying pulumi GCP with GO, I'm trying to set ForceDestroy to bucket but I wasn't able 😞

go
bucket, err := storage.NewBucket(ctx, "my-bucket", &storage.BucketArgs{
			ForceDestroy: true,
		})
		if err != nil {
			return err
		}

Any idea?

I don’t really get what the

registerOutputs()

method is for ? In the example here https://pulumi.io/reference/component-tutorial.html a component property

this.bucketName

is added. In fact we can access the

S3Folder.bucketName

without using

registerOutputs()

so I don’t see the point here ?

Sorry, I have another question… I’m trying to improve my stack’s code now. In my

index.ts

I’m creating first a gke cluster then I’m adding the external dns

Chart

. Problem is, the chart needs the

k8sProvider

from the

Cluster

, or pulumi is running both in parallel, which of course makes the

Chart

accounts and roles creation fail. I tried to add my cluster object to the

dependsOn

key of the

Chart

opts

but it doesn’t wait for it anyway. How can I ensure something is created before anything else depending on it runs ? I thought it would be automatic by referencing the provider from the cluster, but it’s not.

This message was deleted.

Guys, My last question was very lame 😉 I should have asked how can I write Pulumi.Output into a file....what is the best practice? 😉 I have to write a yaml...

Trying again with

golang

,

ctx.Export("bucketName", bucket.Url())

can be exported to a string type to reuse it later?

Maybe a stupid question .. i know there is a tf2pulumi project... is there a pulumi2tf ? I ask because some places insist in their code being TF, but we could write it in pulumi and provide the tf code as an artefact.

Guys, in Python Pulumy-Kubernetes, do you have resource for configmaps? Should I look for this here? https://github.com/pulumi/pulumi-kubernetes/tree/master/sdk/python/pulumi_kubernetes

Sorry for my Stupid question I have just foundit...kubernetes.core.v1 import ConfigMap

Guys in Python what is the best practice to start a kubectl client?

I should set user .kube/config file?

or is there any client I can init with my kubeconfig? or kubeconfigs if I have more then one K8s...so what is the best practice to init the connection?

I got this error: error: Unable to read kubectl config: invalid configuration: no configuration has been provided error: an error occurred while advancing the preview

it's working with the default kubectl setup, but please tell me if there is another way to use this...Thanks

Guys, I would have another question 😉 So I would like to put a configmap resorce into a an EKS, The configmap resource explicitly depends on 2 resources, actually 2 aws role resource....but implicitly it depends on the EKS Cluster itself. How should I code this dependency with Pulumi? because Cluster setup takes longer then the aws role resource creation.

If a

Deployment

fails (no matter the reason) the

Service

associated with it keeps hanging pulumi for hours in the state

Finding Pods to direct traffic to

while it seems it could definitely know there’s nothing left to wait for and should give up immediately. Also, trying to cancel manually in this situation usually ends up in an unrecoverable state where the best way to get out easily from this is to completely

rm

the stack and start over. That’s a big waste of time right now.

Related to my previous message, when creating a new cluster, services to deploy in it have to wait a (very) long time and usually timeout before the cluster is provisioned, forcing to re-run pulumi at least one more time to get everything up and running, that’s not good for any automation process. Is there a way to make them wait longer or to enforce any kind of dependency readiness ?

I am trying to create an EKS cluster and I am running into this error.

aws:eks:Cluster (cluster-eksCluster):
    error: Plan apply failed: error creating EKS Cluster (cluster-eksCluster-f2f0a42): UnsupportedAvailabilityZoneException: Cannot create cluster 'cluster-eksCluster-f2f0a42' because us-east-1e, the targeted availability zone, does not currently have sufficient capacity to support the cluster. Retry and choose from these availability zones: us-east-1a, us-east-1b, us-east-1c, us-east-1d, us-east-1f
    	status code: 400, request id: 25e1928f-0585-11e9-860f-afb328b78912

Where do I set the availability zone?

@chilly-photographer-60932 AZ comes from the subnet

@cuddly-eye-68174 I am using this https://github.com/pulumi/pulumi-eks

I am not setting any subnet

Ugh. I’ve mistakenly created a project & two stacks under my own name, instead of the company’s. Can I move them or should I start again?

@chilly-photographer-60932 where you try to install this? in which AWS region? it's not publicly available everywhere.

Guys, Is it possible to delete a resource from the stack? without dependency check?