pulumi-community #general

general

thankful-optician-22583

10/28/2019, 11:47 PM

Hi, is there a way to replace strings files in pulumi like in terraform. I want to pass in a kubeconfig into a K8s provider and pass in config variables in to the string of the file:

apiVersion: v1
preferences: {}
kind: Config
clusters:
- cluster:
    server: ${server}   
    certificate-authority-data: ${certificate-authority-data}
  name: current-cluster
contexts:
- context:
    cluster: current-cluster
    user: runner-user
  name: runner-context
current-context: runner-context
users:
- name: runner-user
  user:
    token: ${token}

wet-sunset-4939

10/29/2019, 2:35 AM

Hi All, Quick question regarding to Self-service of ServicePrincipal on Azure. Seems pulumi is not support this yet and we may need to develop a DynamicResourceProvider. So I would like to check is there anyone already implement this?

ambitious-helmet-24352

10/29/2019, 4:11 PM

Hey how to I run pulumi up on my stack while excluding certain items ?

better-rainbow-14549

10/29/2019, 5:18 PM

https://github.com/terraform-providers/terraform-provider-azurerm/issues/4747

early-musician-41645

10/29/2019, 6:35 PM

I'm trying to run this code:

const tentacleVpcId = tentacleConfig.require("vpcId");
console.log("VPC-ID: "+tentacleVpcId);

export const tentacleVpc = awsx.ec2.Vpc.fromExistingIds("tentacle-vpc", {
  vpcId: tentacleVpcId,
});

However I get this error:

Error: invocation of aws:ec2/getVpc:getVpc returned an error: invoking aws:ec2/getVpc:getVpc: no matching VPC found

I've checked and rechecked everything (account, credentials, resource ids, config, stack) and still no luck. I ran a

pulumi preview --debug

and I see that there's this output:

Reading AWS VPC: {
  Filters: [{
    Name: "isDefault",
    Values: ["true"]
  }]
}

The VPC is not the default one. It's one of many VPCs. How can I make pulumi get a non-default VPC?

rich-lamp-32782

10/30/2019, 12:57 AM

Does anyone have any examples of editing K8s kube-system configmaps that aren’t being defined in IaC in pulumi? Ideally I’d like to adopt them and then append to them

better-actor-92669

10/30/2019, 9:29 AM

Hello Pulumi Community! Silly question, but how to get the actual value of pulumi output object `<pulumi.output.Output object at 0x107d985b0>`or

<pulumi_gcp.compute.ha_vpn_gateway.HaVpnGateway object at 0x107523430>

in python? I want to assign it to a variable or use it later in a script. Thank you in advance!

witty-yacht-82771

10/30/2019, 11:19 AM

Hey, question about how to check values in testing. I have the following in my `index.ts`:

export const service = new awsx.ecs.FargateService("click-router", {
  assignPublicIp: false,
  cluster,
  subnets: pubSubnets,
  desiredCount: 2,
  taskDefinition: taskDef
});

and then in my tests, I’m trying to do:

it("should have 2 desired instances", async () => {
    const di = await promise(service.service.desiredCount);
    expect(di).to.equal(2);
  });

I get the error that the variable

di

is undefined. How can I do tests on assigned values prior to deploying?

quiet-wolf-18467

10/30/2019, 11:45 AM

OK. I've hit an interesting snag. When GKE releases a new nodeVersion/masterVersion - this causes the dynamic provider to be "replaced" - in-turn replacing all resources that use it 😮

quiet-wolf-18467

10/30/2019, 11:46 AM

pulumi:pulumi:Stack                                  devrel-metrics-production              
 ~   ├─ gcp:container:Cluster                             metrics                    update      [diff: ~minMasterVersion,nodeVersion]
 +-  ├─ pulumi:providers:kubernetes                       metrics                    replace     [diff: ~kubeconfig]
 +-  ├─ kubernetes:core:Namespace                         community                  replace     [diff: ~metadata,provider]
 +-  ├─ kubernetes:core:ServiceAccount                    community                  replace     [diff: ~metadata,provider]
 +-  ├─ kubernetes:<http://rbac.authorization.k8s.io:Role|rbac.authorization.k8s.io:Role>         community                  replace     [diff: ~metadata,provider]
 +-  └─ kubernetes:<http://rbac.authorization.k8s.io:RoleBinding|rbac.authorization.k8s.io:RoleBinding>  community                  replace     [diff: ~metadata,provider,roleRef,subjects]

quiet-wolf-18467

10/30/2019, 11:48 AM

This

provider

is passed along to many smaller projects / repositories - is this going to blow them all away? 😞

quiet-wolf-18467

10/30/2019, 12:17 PM

I took a risk ...

Resources:
    ~ 1 updated
    9 unchanged

quiet-wolf-18467

10/30/2019, 12:17 PM

That was scary

quiet-wolf-18467

10/30/2019, 1:16 PM

@creamy-potato-29402 Is the Helm CLI always required for the Helm stuff? The example states "> NOTE: This example has a dependency on the

helm

CLI. Be sure to install that first! See

instructions below." but I don't can't find anything in the code that would explain why its required

agreeable-australia-82656

10/30/2019, 3:53 PM

any way to dump out the settings on an object so I can see what the values are in order to create them as pulumi code? e.g. I have got a handle to an existing resource: e.g. let mwacl = aws.waf.WebAcl.get( name, id) and would like to examine the values of the contained items

loud-journalist-52263

10/30/2019, 3:54 PM

Hey, folks! Ever since Pulumi 1.{>0} was released, I've been getting the

Ingress .status.loadBalancer field was not updated with a hostname/IP address

as described in https://www.pulumi.com/docs/troubleshooting/#ingress-status-loadbalancer I'm using the Nginx ingress controller and I haven't been able to find any information about what configurations might affect the

status.loadBalancer

field. Does anyone have any tips on resolving this with Nginx? If I revert to 1.0, the timeout goes away.

important-leather-28796

10/30/2019, 3:58 PM

I’m working on a new gcp cluster/stack and have free reign to destroy. I messed up some things and it looks like a cluster was replaced, unfortunately another cluster was lingering so I deleted via the console. Next

pulumi up

referenced something about deleting the old cluster still in progress. Regardless, old cluster is gone, but my stack thinks not. A

pulumi refresh

yields

Preview failed: Get <https://34.70.69.68/api?timeout=32s>: dial tcp 34.70.69.68:443: i/o timeout

. I’m just looking to fully destroy, then move forward again - how can I get synced and destroyed when

refresh

is failing?

ancient-memory-38025

10/30/2019, 4:02 PM

Has anyone got any examples of creating an ECS cluster in one project and then consuming it in another...?? So in project 2 I'm able to use

StackReference

to refer to the ECS cluster created in project 1 but when I run

p up

in project 2 it's generating a a whole new cluster?!

ancient-memory-38025

10/30/2019, 4:03 PM

const tvxClusterS = aws.ecs.Cluster.get(`${identifier}-clustera`, stackCluster.getOutput(config.require("clusterName")))
const tvxVpcIdS = awsx.ec2.Vpc.fromExistingIds(`${identifier}-networka`, {
    vpcId: stackCluster.getOutput(config.require("vpcId"))
})

// const tvxVpcIdS = aws.ec2.Vpc.get(`${identifier}-networka`, stackCluster.getOutput(config.require("vpcId")))

const tvxCluster = new awsx.ecs.Cluster(`${identifier}-clusterb`, {
    cluster: tvxClusterS,
    vpc: tvxVpcIdS
})

const ecsService = new awsx.ecs.EC2Service(`${identifier}-svc`, {
    cluster: tvxCluster,
    taskDefinitionArgs: {
        containers: {
            app: {
                image: awsx.ecs.Image.fromPath(`${identifier}-img`, "./app"),
                portMappings: [{
                    containerPort: 3000,
                    hostPort: 80,
                    protocol: "tcp"
                }]
            }  
        },
    },
    desiredCount: 2
})
const app = new Service(`${identifier}-svc-instance`, ecsService)

ancient-memory-38025

10/30/2019, 4:03 PM

Very stuck 😞

hallowed-pilot-33885

10/30/2019, 4:53 PM

Hello! So when looking at a stack in the UI I can drill into the details of a single resource. Is there a way to do that in the CLI as well?

ambitious-ram-5811

10/30/2019, 5:57 PM

There's no Pulumi API for Azure Bastion right? My cursory

rg

search revealed nada

broad-dog-22463

10/30/2019, 5:58 PM

@ambitious-ram-5811 I believe I released it yesterday in 1.3.1

ambitious-ram-5811

10/30/2019, 5:58 PM

This is extremely timely!

broad-dog-22463

10/30/2019, 5:59 PM

check for

azure.compute.BastionHost

important-leather-28796

10/30/2019, 7:50 PM

We are attempting to create a gcp cluster sandbox for testing out configurations, and pulumi is failing to create a NodePool with

f1-micro

, specifically

invalid machine type fl-micro in zone us-central1-a: googleapi: Error 404: The resource 'projects/acme/zones/us-central1-a/machineTypes/fl-micro' was not found, notFound

, but I can manually create the same thing in the gcp console. What am I missing?

broad-dog-22463

10/30/2019, 7:54 PM

@important-leather-28796 is that a typo? f1 / fl

important-leather-28796

10/30/2019, 7:57 PM

YES! wow, i’m browsing the api looking at instance types and it’s a typo. Thanks Paul.

👍 1

glamorous-television-20475

10/30/2019, 8:00 PM

Is there any recommendation on handling kustomize with pulumi? I saw https://github.com/pulumi/pulumi-kubernetes/issues/370 which appears to be open but fairly stale. It seems like you could just the

kustomize build

command to generate the final resource yaml and use that with

from pulumi_kubernetes.yaml import ConfigFile

ambitious-helmet-24352

10/30/2019, 8:22 PM

hey looking for a typescript safe for istio like is there a specific library ?