pulumi-community #general

chilly-rainbow-79265

08/24/2020, 6:19 PM

Hi, I don't know whether it's already covered or not. I'm trying to create / find a solution, where I will create a cassandra/es cluster with pulumi. The idea is kind of a kubernetes operator, where when I scale, the nodes will be added to the cluster(including the configuration and all). I was reading through the docs and found out that there is something called component and using that we can have abstraction. But is it possible to have methods for componants ?

Copy code

type Cassandra struct {
    pulumi.ResourceState
}

func NewCassandra(ctx *pulumi.Context, name string, opts ...pulumi.ResourceOption) (*MyComponent, error) {
    cassandra := &Cassandra{}
    err := ctx.RegisterComponentResource("pkg:index:NewCassandra", name, cassandra, opts...)
    if err != nil {
        return nil, err
    }

    return cassandra, nil
}

Here is it possible to some how, I can add a method for

Copy code

cassandra.Scale()

or more preferably, when I increase the node from 1-2 it'll get added to the cluster, that means, the state change event from the engine, should it be possible to do something there. Main thing I want is the capability to plug the scale method with increase/decrease number of nodes. ie, if I change the number of nod*es from 1 -> 2 it should trigger a scale up* op, and if 2-> 1 it should do scale down + destroy the vm.

stocky-spoon-28903

08/24/2020, 6:23 PM

Copy code

func (c *Cassandra) AMethod(/* params */) error {
    // ...
}

stocky-spoon-28903

08/24/2020, 6:24 PM

(You also probably want your

NewCassandra

function to return

*Cassandra

rather than

*MyComponent

chilly-magazine-6129

08/24/2020, 6:24 PM

How can I get the uploaded lambda .zip file that Pulumi generates? I'd like to run it locally with some of the packages that simulate lambdas offline

damp-secretary-81951

08/24/2020, 8:00 PM

What’s the process for sharing a Pulumi stack w/ someone else on my team? 1. Add them as a user to my Pulumi org 2. Have them pull down &

cd

into our git repo 3. Will

pulumi up

work at that point? I can only find docs for initializing a new stack, not starting to use an existing one

hundreds-musician-51496

08/24/2020, 8:38 PM

Regarding the StackReference resource - how do I ensure the output from a given reference is available? Does pulumi offer a way to run

up

on a stack and all its dependencies (i.e., all the stacks that it references as resources)?

wet-noon-14291

08/24/2020, 9:22 PM

One thing that I have seen on twitter as well as I've been asked myself while trying to promote pulumi is the concern around how up to date pulumi is? If something changes in azure (as it is in my case), how fast will pulumi be able to react and add support? It is a valid concern I would say, but I don't think it is a major issue. It would be nice with an official post addressing that explaining why it most likely isn't a concern.

few-shoe-6011

08/24/2020, 9:44 PM

Hi, I'm using pulumi/kubernetes in nodejs and I'm a little confused why the deployment name is showing up as

[secret]

. Any idea? I am using a secret from pulumi to populate a kubernetes secret that is eventually mounted as a volume on the deployment, but I don't understand why that would cause the name to be masked.

Copy code

const envSecret = new kx.Secret(`${env}-api-secret`, {
  stringData: {
    ...
    GOOGLE_APPLICATION_CREDENTIALS: config.requireSecret('GOOGLE_APPLICATION_CREDENTIALS'),
  }
});
...
export const deployment_name = deployment.metadata.name;

damp-secretary-81951

08/24/2020, 11:30 PM

Is there any way to specify a different entrypoint for a particular stack? For instance, I want my dev, staging & prod stacks to all use

index.js

, but my

quick-demo

stack to use

quick-demo.js

instead.

nice-football-42636

08/25/2020, 10:50 AM

Hi guys. I'm working with the

.NET SDK

and I would like to know, if there is a way to access the current stack name (selected via

pulumi cli

) before running the deployment. I found in the source code a property

Deployment.Stack

but it throws an exception if I try to access it before running the deployment. I need to do that in order to run deployment of 2 stacks in one azure dev ops pipeline through one project (we use monolithic approach to project/stacks structure). So I would like to access the currently selected stack and based on that run deployment of specific stack from code. (I know I can access it via

PULUMI_STACK

environment variable but I don't want to rely on some deep internal implementation details but rather use some top level api to access the stack name).

straight-branch-26699

08/25/2020, 1:47 PM

Hi Team, We are looking at provisioning AWS resources. Is there a way to pick aws cli profiles from Pulumi config files for the same

gifted-cat-21399

08/25/2020, 4:59 PM

Hello! Having issues trying to run Pulumi in Azure Pipelines. Our Pulumi project has a reference to a nuget package that we call "PulumiFactory" that sits in our private nuget repository on an installation of Artifactory. We can run Pulumi up locally but not in the Azure Pipeline. In Azure, we are using the Pulumi extension, hence Pulumi task. We have access to our Artifactory repository of nuget packages. I've retrieved the necessary nuget package successfully using a variety of methods in a task before attempting the Pulumi task in the pipeline. The baked-in Pulumi build script still seeks the "missing" package via nuget.org and obviously cannot find it as it's in another source. We wish not to abandon our approach of using this custom Pulumi Factory we've built and reference in the project. The error in the pipeline is basically the following: D:\a\1\s\DevOpsDevTest\XXX-USC-DEV-RG\XXX-USC-DEV-RG.csproj : error NU1101: Unable to find package OurCompany.PulumiFactory. No packages exist with this id in source(s): Microsoft Visual Studio Offline Packages, nuget.org. Can anyone help with this? Pulumi in Azure Pipelines, Pulumi referencing custom nuget package in our own Artifactory.

many-psychiatrist-74327

08/25/2020, 9:15 PM

👋 Hello! I’m relatively new to Pulumi and have a presumably basic question: how do I get the kubeconfig file after creating a k8s cluster in EKS (in Python)? I want to: (1) create an EKS cluster, and then (2) deploy an app onto it. For (2), I need to pass the cluster’s information, otherwise I get an error of the form:

unable to load Kubernetes client configuration from kubeconfig file

. I see that in Crosswalk for AWS, the

@pulumi/eks

(nodejs) package would let me export

cluster.kubeconfig

(docs). But I don’t see anything similar for Python! And the Cluster Resource itself doesn’t seem to have any Output that has the necessary information to set my

kubeconfig

or otherwise create a Provider to use in my Kubernetes resources. Please help!

handsome-zebra-11018

08/26/2020, 2:45 AM

I am looking to attach an existing nfs volume to container in EKS cluster. The NFS volume already attached to all nodes on EKS cluster and wanted to attach it with container. Can anyone recommend how it can be done using pulumi?

Copy code

containers: [
              {
                name: appServerLabel,
                image: config.get("docker_image") + ":" + config.get("docker_image_tag"),
                ports: [{name: "http", containerPort: 80},{name: "https", containerPort: 443}],
                volumeMounts: [
                  {
                      name: "volume-bigdisk",
                      mountPath: "/logs/"
                  }
                ],
              }
           ],
volumes: [
              {
                name: "volume-bigdisk",
                server: "servername*****",
                path: "/af_appserver/VAF/bigdisk",
                readOnly: false
              }
         ]

I am getting below error

Copy code

ame: string; server: string; path: string; readOnly: boolean; }[]; }' is not assignable to type 'PodSpec'.
                    Types of property 'volumes' are incompatible.
                      Type '{ name: string; server: string; path: string; readOnly: boolean; }[]' is not assignable to type 'Input<Input<Volume>[]>'.
                        Type '{ name: string; server: string; path: string; readOnly: boolean; }[]' is not assignable to type 'Input<Volume>[]'.
                          Type '{ name: string; server: string; path: string; readOnly: boolean; }' is not assignable to type 'Input<Volume>'.
                            Object literal may only specify known properties, and 'server' does not exist in type 'Input<Volume>'.

breezy-butcher-78604

08/26/2020, 4:38 AM

the documentation for

apply()

says that the function cannot create resources, however i have a template that is creating resources within an

apply()

and it appears to work ok... is there a reason i shouldn't do this?

incalculable-dream-27508

08/26/2020, 10:50 AM

Is there a way to manage machines in two independent OpenStack clusters "at once"?

hallowed-barista-41250

08/26/2020, 1:18 PM

Hey all, just discovered Pulumi. Am I correct in thinking if I have a PHP app, I can still use Pulumi even though it isn't a supported language? That's just for writing the configuration?

icy-jordan-58549

08/26/2020, 2:59 PM

is there any way to get

pulumi stack output

from the code? I was pretty sure that I will get stack output from Webhook, but it seems to be was not planned. Any suggestions?

handsome-zebra-11018

08/26/2020, 3:15 PM

Anyone has example how to use existing NFS Volumes with pulumi ?

chilly-rainbow-79265

08/26/2020, 4:32 PM

Is there any way for

life cycle hooks

. I know it's still in progress. But is there

any way for chaining an action

, after a resource got created ?

cool-telephone-9886

08/26/2020, 5:28 PM

Slack seems like the wrong forum for asking questions - just screaming into the wind here but: I can’t get pulumi to create an azure aks cluster with a virtual node. I’ve used the AddonProfile with AciConnectorLinuxArgs but no virtual node is added to the cluster once it completes being created. What am I missing? Thanks.

damp-secretary-81951

08/26/2020, 5:34 PM

I’m currently updating an S3 bucket (HEAD, PUT) from an API Gateway lambda. I heard that if you use VPC, the requests will be faster than if it connects via a public IP. Do I need a VPC endpoint for this (

aws.ec2.VpcEndpoint

) or does Pulumi set this up automatically?

adamant-advantage-95831

08/26/2020, 7:22 PM

Has anyone had success using

pulumi stack change-secrets-provider <secrets-provider>

bitter-application-91815

08/26/2020, 7:32 PM

@gentle-diamond-70147 hey there

bitter-application-91815

08/26/2020, 7:33 PM

Have a few questions as to how to pull out the vpc and security details from a aws clusterInstance

bitter-application-91815

08/26/2020, 7:34 PM

and then assign them to a lambda (which i suspect is through a resource ?)

bitter-application-91815

08/26/2020, 7:37 PM

once i can validate all the cluster stuff, i'll move the account over to a paid one, my boss has OK'd it.

millions-judge-24978

08/26/2020, 10:11 PM

Has anyone managed to use https://github.com/actions/cache to cache the pulumi plugins?

little-cartoon-10569

08/27/2020, 3:38 AM

I have made an oopsie with my changes. Pulumi has begun to replace my VPC with a new one, but has failed because of permissions, dependencies, and more. Not worried about that, I can sort it. However, I don't want to replace my VPC. Is there anything I can do to get Pulumi to not want to finish deleting it each time I

pulumi up

😱 1

little-cartoon-10569

08/27/2020, 4:39 AM

New problem. Is it possible to change the provider on a resource (in my case, a VPC) without replacing the resource? I don't want to delete-create, but I want to go from using the default (undefined) provider to an explicitly-defined provider that is pointing to the same AWS account and profile.