pulumi-community #kubernetes

gorgeous-elephant-23271

04/14/2020, 8:07 PM

I've tested the creds locally and they work

eager-pillow-75917

04/22/2020, 12:46 PM

hi, I want to deploy a chart from a repo, not local, but I want to use a values.yaml file to replace some values. How do I do that?

bland-lamp-16797

04/24/2020, 2:45 PM

I've notice that pulumi does not recreate missing manifest... like, if I create namespace and namespace is gone, it will not recreate

bland-lamp-16797

04/24/2020, 2:46 PM

is it same for other k8s resources?

busy-account-42432

04/26/2020, 10:03 AM

Is there a PR for renderyamltodirectory option in the kubernetes provider for golang? I'm new to pulumi and this would be required for our GitOps workflow

gorgeous-animal-95046

04/27/2020, 7:36 PM

is it not possible to

import

CustomResource

? The diff is always different because the

spec

doesn't match up to what's in my program (the program has the same

spec

as what's currently in k8s). The diff when running

pulumi up

has a complete

spec

inside of it, like the remote was empty. I'm using dry run for importing.

bland-lamp-16797

04/29/2020, 3:15 PM

did anyone install nginx-ingress in pulumi ? I'm looking at examples but it might be someone did already this before me

cool-family-11574

05/01/2020, 8:30 AM

Is there a way to convert yamlto Pulumi?

better-rainbow-14549

05/01/2020, 8:31 AM

not that i know of, but you can convert yaml to json and then use that as the majority of your spec block

billowy-army-68599

05/01/2020, 3:42 PM

@cool-family-11574 sorry for the delay! You can't convert yaml to pulumi per se, but we do have ConfigFile https://www.pulumi.com/docs/reference/pkg/nodejs/pulumi/kubernetes/yaml/#ConfigFile This'll create a set of resources from a YAML config (remote or local) and you can also do transformations on it, it's just an object/dict. Here's an example

Copy code

const crds = new k8s.yaml.ConfigFile("crds", {
    file: "<https://github.com/jetstack/cert-manager/releases/download/v0.14.1/cert-manager.crds.yaml>",
}, { provider: provider })

cool-family-11574

05/01/2020, 3:57 PM

Is this only available in Node JS or other languages

creamy-potato-29402

05/01/2020, 4:03 PM

https://github.com/itowlson/vscode-pulumi-kubernetes

creamy-potato-29402

05/01/2020, 4:04 PM

kind of. it's a vscode extension though.

hundreds-portugal-17080

05/05/2020, 2:23 AM

Hello, Is there a way to restart a container inside pod from pulumi code. I am using postStart hook to install some libraries but need a container restart to make that into effect. I could have precompiled the docker image with set of liraries but those are too many and so, they are passed as pulumi config parameters to pulumi codebase.

busy-umbrella-36067

05/05/2020, 11:11 PM

how hard would it be to build caching into

k8s.yaml.ConfigFile

if the

args.file

is a URL? I’m using one in a module and its making 16 concurrent requests to Github releases which gets ratelimited for obvious reasons.

hundreds-portugal-17080

05/06/2020, 4:34 AM

Hello, I am trying to override a multiline value(filebeatConfig) from https://github.com/elastic/helm-charts/blob/master/filebeat/values.yaml through pulumi code and it just doesn't work. I get "Error: 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" Attached chart code as an example. Not sure if I am doing something wrong. I have very basic knowledge on helm. Any help is appreciated.

Untitled

hundreds-receptionist-31352

05/06/2020, 5:10 PM

Hi could anyone get working this -> https://www.pulumi.com/docs/guides/adopting/from_kubernetes/

hundreds-receptionist-31352

05/06/2020, 5:11 PM

Copy code

import * as k8s from "@pulumi/kubernetes";
// Deploy the latest version of the stable/wordpress chart.
const wordpress = new k8s.helm.v2.Chart("wpdev", {
    repo: "stable",
    chart: "wordpress",
    version: "2.1.3",
});
// Export the public IP for WordPress.
const frontend = wordpress.getResource("v1/Service", "wpdev-wordpress");
export const frontendIp = frontend.status.loadBalancer.ingress[0].ip;

hundreds-receptionist-31352

05/06/2020, 5:12 PM

just following the same steps It never up , a lot of pods are waiting for some extra variables to run properly

polite-motherboard-78438

05/07/2020, 6:20 PM

hello. I want to use the Postgres provider to manage databases on a Postgres instance deployed on Kubernetes. The issue is that the service is not exposed to the outside so Pulumi can´t reach it. What is the best way to handle this? Should I use a VPN/kubectl port-forwarding? I need a way that can be use in CI envrionment and also for new/existing clusters. (A new cluster will create the K8s cluster, deploy the postgres StatefulSet and setup default databases).

abundant-airplane-93796

05/08/2020, 12:51 PM

is there any way to enforce that a provider MUST be specifically sent for any/all kubernetes resources? on a couple of occasions I've inadvertantly created resources on whatever cluster my local kubectl context is set to instead of one of those my Pulumi code is managing - basically by forgetting to add the 'provider' setting appropriately

future-air-70511

05/08/2020, 2:47 PM

I'm incrementally trying get sonarqube deployed using their helm chart via pulumi/c#. When I supply the values in a yaml file with hardcoded values on the command line, it works successfully. When I try to use pulumi, with the following code I receive the following error:

Copy code

new ChartArgs
{
    Repo = "oteemocharts",
    Chart = "sonarqube",
    Namespace = namespaceName,

    Values =
        {
            {  "image", new { tag = "8.3-community" } },
            {
                "postgresql",
                new
                {
                    enabled = false,
                    postgresqlServer = postgresqlServer.Fqdn,
                    postgresqlPassword = postgresqlPassword.Result,
                    postgresqlDatabase = postgresqlDatabase.Name,
                    postgresqlUsername = Output.Format( $"{postgresqlServer.AdministratorLogin}@{postgresqlServer.Name}" )
                }
            },

Copy code

Diagnostics:
  kubernetes:helm.sh:Chart (helm-sonarqube):
    error: Pulumi.ResourceException: coalesce.go:195: warning: destination for postgresqlDatabase is a table. Ignoring non-table value sonarDB
    coalesce.go:195: warning: destination for postgresqlPassword is a table. Ignoring non-table value sonarPass
    coalesce.go:195: warning: destination for postgresqlDatabase is a table. Ignoring non-table value sonarDB
    coalesce.go:195: warning: destination for postgresqlPassword is a table. Ignoring non-table value sonarPass
    Error: template: sonarqube/templates/secret.yaml:14:66: executing "sonarqube/templates/secret.yaml" at <b64enc>: wrong type for value; expected string; got map[string]interface {}

Anyone have ideas what I'm doing wrong?

brash-manchester-88595

05/08/2020, 3:44 PM

Is there a recommended way to refactor the resource hierarchy when one of the resources that moves is a provider? Specifically, try moving around an

eks.Cluster

(eg. from top level into a component resource). It’ll work almost, (with

aliases: [{parent: pulumi.rootStackResource}]

), but you can’t move the

nodeAccess

configmap, because it can’t find it’s provider (it’s trying to use the new provider that you don’t have yet). I can open a bug if this isn’t tracked yet, I ran into problems with refactoring providers multiple times already. Possibly the problem is that the new provider doesn’t act like a dependency on which you have to wait. Or that the reference to the provider itself isn’t aliased (ie. the alias for the provider isn’t propagated to the dependents so that you could reference the old provider through the new urn).

brash-manchester-88595

05/08/2020, 4:40 PM

vaguely remembered the bug, here it is: https://github.com/pulumi/pulumi/issues/3979

gorgeous-elephant-23271

05/10/2020, 1:15 PM

Hi there, I'm noticing that old replica sets aren't being cleaned up after a successful deployment. I think it kinda makes sense as you'd want to be able to roll back to and old one on failure, but is there a standard way of cleaning these up?

ancient-megabyte-79588

05/13/2020, 8:54 PM

Hello again Kubernetes peeps...

ancient-megabyte-79588

05/13/2020, 8:56 PM

This fragment of Pulumi-code that uses the nginx Ingress Controller Helm chart is working but it does not use my k8sDnsLabel in the PublicIP that is created in Azure by the IngressController. I'm wondering if any of you have any insights.

Copy code

const env = pulumi.getStack(); // reference to this stack
const stackId = `dave/aks/${env}`;
const aksStack = new pulumi.StackReference(stackId);
const k8sDnsName = aksStack.getOutput("k8sDnsName"); // <-- This is "identity-auth-dev"

// Deploy ingress-controller using helm to AKS Cluster
const options = {
    chart: "nginx-ingress-controller",
    namespace: "kube-system",
    repo: "bitnami",
    values: {
        annotations: {
            "<http://service.beta.kubernetes.io/azure-dns-label-name|service.beta.kubernetes.io/azure-dns-label-name>": "identity-auth-dev"
        },
        resources: { requests : {memory: "150Mi", cpu: "100m"}},
        serviceType: "LoadBalancer",
        nodeCount: 1,
    }
};
const nginxIngress = new k8s.helm.v3.Chart("nginx", options, {provider: k8sProvider });

ancient-megabyte-79588

05/13/2020, 8:56 PM

I don't think this a Pulumi problem other than me doing something incorrectly.

kind-mechanic-53546

05/16/2020, 6:53 AM

has anyone here used the pulumi crosswalk stacks Supposed to be production ready, maybe best practice?

faint-motherboard-95438

05/18/2020, 9:31 PM

Hello there, I have a weird issue with a chart, I’m using https://github.com/bitnami/charts/tree/master/bitnami/postgresql-ha, everything works fine so far, but each time I

pulumi up

Secret

created by the chart (namely the

pgpool

one) keeps getting replaced each time for a

~data

diff reason. Anyone would know the reason and how to make it permanent please ? (Never had this issue with any other chart I use)