pulumi-community #kubernetes

quiet-state-42882

06/07/2020, 6:04 AM

I seem to be experiencing issues in deploying a helm chart. The issues are as follows: * resource drone-drone was successfully created, but the Kubernetes API server reported that it failed to fully initialize or become live: 'drone-drone' timed out waiting to be Ready * Service does not target any Pods. Selected Pods may not be ready, or field '.spec.selector' may not match labels on any Pods Am I missing something in the configuration? Or would this suggest a misconfiguration in the helm chart?

famous-jelly-72366

06/09/2020, 2:02 PM

anyone also getting warning from cert-manager crds about

<http://apiextensions.k8s.io/v1beta1/CustomResourceDefinition|apiextensions.k8s.io/v1beta1/CustomResourceDefinition>

being deprecated? I tried doing a helm chart transform to fix this, but seems like the cert-manager webhook doesn't work with the updated v1 :S

famous-jelly-72366

06/10/2020, 9:51 AM

Is it expected behavior that upgrading an AKS cluster will replace all resources ???

bland-lamp-16797

06/10/2020, 3:11 PM

is there a way to install helm with pulumi and still having it when i type

helm list

bland-lamp-16797

06/10/2020, 3:12 PM

or to simply change

<http://app.kubernetes.io/managed-by|app.kubernetes.io/managed-by>: pulumi

somehow

bland-lamp-16797

06/10/2020, 3:18 PM

label validation error: key "app.kubernetes.io/managed-by" must equal "Helm": current value is "pulumi";

bland-lamp-16797

06/10/2020, 3:21 PM

also, after upgrading GCP cluster, I'm getting this error when trying to run helm

error: apiVersion "extensions/v1beta1/Deployment" was removed in Kubernetes 1.16. Use "apps/v1/Deployment" instead.

full-dress-10026

06/10/2020, 4:52 PM

I'm switching from

new aws.eks.NodeGroup()

eksCluster.createNodeGroup

and noticing that

createNodeGroup

does not include the

nodeRoleArn

property in the args object. How would I provide an IAM role arn to a node group created with the

createNodeGroup

method?

full-dress-10026

06/10/2020, 5:19 PM

Wait, I think I have thoroughly confused myself. What is the proper way to create a managed node group in Pulumi? There appears to be many ways and it is unclear which blog posts are out of date.

famous-jelly-72366

06/11/2020, 11:28 AM

how are people handling deploying apps to k8s? do the pulumi code for the app live in the apps repo (as a separate stack) ? and then CI uses this to deploy to k8s, handling e.g. different branches etc.

miniature-rose-15269

06/11/2020, 2:22 PM

is it possible to wait for a cert-manager certificate secret to have the cert keys filled in? how would one go about it?

limited-rainbow-51650

06/11/2020, 2:28 PM

Pulumi can render YAML from code (https://www.pulumi.com/blog/kubernetes-yaml-generation/) but can it also do the reverse? I give it YAML and it gives me some Typescript on which I can continue working?

famous-jelly-72366

06/12/2020, 12:22 PM

when using helm charts with CRDs (e.g. cert-manager) it looks like those resources' urn aren't properly prefixed. At least I'm getting errors when trying to bring up a second cluster also using cert-manager claiin that the urns are not unique :S

Copy code

kubernetes:<http://apiextensions.k8s.io:CustomResourceDefinition|apiextensions.k8s.io:CustomResourceDefinition> (<http://certificaterequests.cert-manager.io|certificaterequests.cert-manager.io>):
    error: Duplicate resource URN 'urn:pulumi:k8s::tms-pulumi::kubernetes:<http://helm.sh/v2:Chart$kubernetes:apiextensions.k8s.io/v1beta1:CustomResourceDefino/v1beta1:CustomResourceDefinition::certificaterequests.cert-manager.io';|helm.sh/v2:Chart$kubernetes:apiextensions.k8s.io/v1beta1:CustomResourceDefino/v1beta1:CustomResourceDefinition::certificaterequests.cert-manager.io';> try giving it a unique name

limited-rainbow-51650

06/12/2020, 2:57 PM

In a Pod spec, how should I convert a command with an input pipe like below into the TS counterpart?

Copy code

command:
  - bash
  - -ec
  - |
      # Execute entrypoint as usual after obtaining ZOO_SERVER_ID based on POD hostname
      HOSTNAME=`hostname -s`
      if [[ $HOSTNAME =~ (.*)-([0-9]+)$ ]]; then
        ORD=${BASH_REMATCH[2]}
        export ZOO_SERVER_ID=$((ORD+1))
      else
        echo "Failed to get index from hostname $HOST"
        exit 1
      fi
      exec /entrypoint.sh /run.sh

salmon-ghost-86211

06/12/2020, 3:48 PM

How would I enable group metrics collection for a K8s node group autoscaling group in AWS? Group metrics are not collected by default and when I check the Monitoring tab on the ASG it has a link

Enable Group Metrics Collection

. I want it enabled when it is created. I am using the

eks

APIs and calling

eks.Cluster

and then

cluster.createNodeGroup

. I have been able to retrieve from the nodeGroup properties the actual ASG name, but I am not sure how to either update an ASG to have the metrics enabled or to enable metrics on nodeGroup creation.

gorgeous-elephant-23271

06/13/2020, 9:33 PM

Hi, if I delete an entire k8s cluster and have other stacks that depend on it, is there a clean way to delete that stack entirely? It looks like with

pulumi state delete <urn>

I need to recurse through each resource and delete leaves first

gorgeous-elephant-23271

06/13/2020, 9:42 PM

running

pulumi stack rm <stack> --force

works, but feels a little ugly as I'm recreating the entire thing

calm-farmer-19058

06/18/2020, 7:42 PM

Hey all, have been getting my aws-typescript stacks setup and running in github actions, the stack is trying to create a kubernetes deployment. Kubernetes authentication is failing on eks get-token as defined in the kubeconfig, pulumi is not using the aws-region defined in the stack config so fails to connect to the cluster. By adding AWS_DEFAULT_REGION environment variable then everything is fine, this seems like a bug or am I missing something?

hundreds-portugal-17080

06/18/2020, 10:20 PM

Hello, When I teardown, some of the load balancers created from helm charts is not cleaning up. It is not happening every time and highly inconsistent. Please advice.

busy-soccer-65968

06/19/2020, 9:58 PM

https://pulumi-community.slack.com/archives/C84L4E3N1/p1592603831244400

salmon-ghost-86211

06/22/2020, 3:59 PM

Bumping this since I still need an answer https://pulumi-community.slack.com/archives/CRFURDVQB/p1591976889392000

faint-motherboard-95438

06/23/2020, 2:32 PM

Hi there, Trying to use

getResource()

on a

k8s.helm.v2.Chart

yields a weird error:

Error: invocation of kubernetes:yaml:decode returned an error: error converting YAML to JSON: yaml: line 29: could not find expected ':'

I’m installing

mongodb-replicaset

chart and try to access the

StatefulSet

Copy code

this.statefulSet = this.chart.getResource(
      'apps/v1/StatefulSet',
      replicaSetName,
    )

    const uri = `mongodb://${this.statefulSet.spec.serviceName}:27017`

Following snippet example from here : https://www.pulumi.com/docs/guides/adopting/from_kubernetes/#provisioning-a-helm-chart

icy-jordan-58549

06/25/2020, 4:19 PM

Hi, this code fails with error:

Preview failed: resource operator/kafka-bootstrap-lb does not exist

Copy code

export const kafkaService = k8s.core.v1.Service.get(
  "kafkaLB",
  "operator/kafka-bootstrap-lb"
);
const kafkaRecord = new azure.dns.ARecord(
  "kafka",
  {
    name: "kafka",
    zoneName: zone.name,
    resourceGroupName: config.resourceGroup.name,
    ttl: 60,
    records: [kafkaService.status.loadBalancer.ingress[0].ip],
  },
  {
    dependsOn: kafkaService,
  }
);

Service

operator/kafka-bootstrap-lb

isn’t presented in resources (helm), due to operator that deploys this later and looks like

pulumi

doesn’t wait for that service.

dazzling-sundown-39670

06/30/2020, 11:17 AM

How can I replicate this command?

Copy code

helm install --name fission --namespace fission \ <https://github.com/fission/fission/releases/download/1.10.0/fission-all-1.10.0.tgz>

I figured fetchOpts.repo but I'm not sure what to put in it

hundreds-portugal-17080

07/03/2020, 12:23 AM

Hello, I am trying to override

filebeatConfig

config with backticks using following code and it doesn't work. I used debug in pulumi up command and it doesn't show much. Errors: template: filebeat/templates/daemonset.yaml2927: executing "filebeat/templates/daemonset.yaml" at <include (print .Template.BasePath "/configmap.yaml") .>: error calling include: template: filebeat/templates/configmap.yaml1335: executing "filebeat/templates/configmap.yaml" at <.Values.filebeatConfig>: range can't iterate over filebeat.yml Chart used: https://github.com/elastic/helm-charts/blob/master/filebeat/values.yaml Pulumi specific code using the chart:

getFilebeatChart(elkChart: k8s.helm.v2.Chart, kibanaChart: k8s.helm.v2.Chart): k8s.helm.v2.Chart | undefined {

if (!this.enabled) {

return undefined;

const filebeatVersion = this.getConfig("filebeatVersion");

const fileBeatConfiguration = this.getConfig("fileBeatConfiguration");

return new k8s.helm.v2.Chart("filebeat", {

path: "../helm_packages_v1/elastic-helm-charts-7.7.0/filebeat",

transformations: [obj => {

if (obj.kind === "DeamonSet" ) {

obj.metadata.annotations=  {"<http://pulumi.com/timeoutSeconds|pulumi.com/timeoutSeconds>": this.esTimeout}

}],

values: {

imageTag: this.esVersion,

filebeatConfig:

``filebeat.yml: |`

filebeat.inputs:

- type: docker

containers.ids:

- '*'

processors:

- add_kubernetes_metadata:

in_cluster: true

Copy code

}, { dependsOn: [ elkChart, kibanaChart], providers: { "kubernetes": this.cluster.provider } ,customTimeouts: {`

create: "2m",

delete: "2m",

update: "2m",

}});

bored-river-53178

07/03/2020, 7:31 AM

does anybody know why k8s.helm.v2.Chart destroys its resources one by one, not in parallel and if that could be fixed? it takes much more time than needed with many manifests in the helm chart

bitter-tiger-55434

07/03/2020, 1:31 PM

Hi guys, I was tring to use

pulumi refresh

to sync k8s cluster's status to local stack. But when I tried to apply the previous code by using

pulumi up

, it tells me there is no updates. Is there anything I misunderstand about

pulumi refresh

nutritious-judge-27316

07/07/2020, 3:16 PM

hi all, looking added apiversion traefik.containo.us/v1alpha1 can someone point me right directions

dazzling-sundown-39670

07/07/2020, 6:10 PM

How can I extend this package (typescript)? I want to add Issuer

most-spoon-17568

07/07/2020, 6:15 PM

I am trying to create eks with node group(eks.nodeGroup sdk). eks version used: 1.16 It creates eks control plane but not the ec2. Pulumi console shows that k8 pods are not able to find nodes and times out. I couldn't find any erors on creation of nodes in pulumi console output. Note: It works perfectly with eks 1.15 with no issues. Are there any compatibility version issues that I need to be aware of? Sdk versions used: "@pulumi/aws": "^2.9.1", "@pulumi/awsx": "^0.20.0", "@pulumi/eks": "^0.19.2", "@pulumi/kubernetes": "^2.3.0", "@pulumi/pulumi": "^2.4.0",