pulumi-community #general

general

best-hydrogen-40271

10/22/2019, 9:31 PM

hmm.

pulumi up

is trying to connect to a cluster that doesnt exist, because it hasn't created it yet

best-hydrogen-40271

10/22/2019, 9:31 PM

specifically when I'm defining a Namespace

best-hydrogen-40271

10/22/2019, 9:36 PM

ah, no provider

best-hydrogen-40271

10/22/2019, 10:11 PM

it's replacing my k8s cluster every time

best-hydrogen-40271

10/22/2019, 10:11 PM

https://pastebin.com/jN1mEGMJ

best-hydrogen-40271

10/22/2019, 10:11 PM

What am I doing wrong?

best-hydrogen-40271

10/22/2019, 10:11 PM

is it the

random.RandomPassword

best-hydrogen-40271

10/22/2019, 10:17 PM

hmm, apparently related to the

nodeconfig

best-hydrogen-40271

10/22/2019, 10:17 PM

but also, seems to fail to actually create the nodepool for the cluster

best-hydrogen-40271

10/22/2019, 10:20 PM

const primary = new GkeCluster("Primary", {});
const primaryStandardNodes = new gcp.container.NodePool(
    "primary-standard-nodes",
    {
        cluster: primary.cluster.name,
        nodeConfig: {
            machineType: "n1-standard-1",
            metadata: {
                "disable-legacy-endpoints": "true"
            },
            oauthScopes: [
                "<https://www.googleapis.com/auth/logging.write>",
                "<https://www.googleapis.com/auth/monitoring>"
            ],
            preemptible: false
        },
        nodeCount: 1
    }
);

best-hydrogen-40271

10/22/2019, 10:22 PM

ok, something was preventing it from completeing the cluster deployment properly if one of the other resources failed

best-hydrogen-40271

10/22/2019, 10:22 PM

looks like it's working now

best-hydrogen-40271

10/22/2019, 10:22 PM

odd behaviour though

best-hydrogen-40271

10/22/2019, 10:23 PM

I need to create a GCS bucket with a specific name

best-hydrogen-40271

10/22/2019, 10:24 PM

pulumi likes tacking on little sha snippets, is there a way around that?

best-hydrogen-40271

10/22/2019, 10:25 PM

eugh, it's still trying to replace perfectly good clusters

best-hydrogen-40271

10/22/2019, 10:25 PM

+-  │  ├─ gcp:container:Cluster                      primary                 replace     [diff: +nodeConfig]

best-hydrogen-40271

10/22/2019, 10:38 PM

is it possible to refernce a resource that I don't want pulumi to manage? e.g. I don't want it to create or destroy it?

best-hydrogen-40271

10/22/2019, 10:39 PM

or is there a way to config pulumi to ignore it on destroy (e.g. a bucket with stuff I want to leave around and re-attach to at a later date)

best-hydrogen-40271

10/22/2019, 10:43 PM

Thanks for all the support 👍 learnt a lot in a short period of time!

best-hydrogen-40271

10/22/2019, 10:44 PM

any resources on best practices for laying out projects?

crooked-jelly-50877

10/22/2019, 10:49 PM

Running into an issue trying to deploy the stable/nginx-ingress helm chart.

crooked-jelly-50877

10/22/2019, 10:49 PM

I’m not sure if this is a pulumi issue, a problem with our TS definition, or a problem with the chart.

crooked-jelly-50877

10/22/2019, 10:50 PM

The issue is simple -but I dont know how to fix it 😉 The

replicaCount

is getting set to a float

1.0

- not an integer

crooked-jelly-50877

10/22/2019, 10:50 PM

This causes the helm template to blow up

crooked-jelly-50877

10/22/2019, 10:50 PM

Sounds like a js/ts casting issue?

crooked-jelly-50877

10/22/2019, 10:51 PM

The error trace is

Error: render error in "nginx-ingress/templates/default-backend-poddisruptionbudget.yaml": template: nginx-ingress/templates/default-backend-poddisruptionbudget.yaml:1:7: executing "nginx-ingress/templates/default-backend-poddisruptionbudget.yaml" at <gt .Values.defaultBackend.replicaCount 1.0>: error calling gt: incompatible types for comparison
    Error: Error: Command failed: helm template /var/folders/br/vk7vq4s17z50g66njs1gmlc80000gn/T/tmp-78327rwjsRXZ39Xkh/nginx-ingress --name nginx-ingress --values /var/folders/br/vk7vq4s17z50g66njs1gmlc80000gn/T/tmp-78327rwjsRXZ39Xkh/nginx-ingress/values.yaml --values /var/folders/br/vk7vq4s17z50g66njs1gmlc80000gn/T/tmp-783273FSDGnkfbySQ.yaml --namespace nginx
    Error: render error in "nginx-ingress/templates/default-backend-poddisruptionbudget.yaml": template: nginx-ingress/templates/default-backend-poddisruptionbudget.yaml:1:7: executing "nginx-ingress/templates/default-backend-poddisruptionbudget.yaml" at <gt .Values.defaultBackend.replicaCount 1.0>: error calling gt: incompatible types for comparison
        at /Users/warren.strange/tmp/forgeops/cluster/pulumi/node_modules/@pulumi/kubernetes/helm/v2/helm.js:112:23
        at OutputImpl.<anonymous> (/Users/warren.strange/tmp/forgeops/cluster/pulumi/node_modules/@pulumi/pulumi/output.js:110:47)
        at Generator.next (<anonymous>)
        at fulfilled (/Users/warren.strange/tmp/forgeops/cluster/pulumi/node_modules/@pulumi/pulumi/output.js:18:58)

crooked-jelly-50877

10/22/2019, 10:51 PM

This is what our defautlBackend ts snippet looks like:

crooked-jelly-50877

10/22/2019, 10:52 PM

defaultBackend: {
                    enabled: "false",
                    replicaCount: 1, // to make helm happy
                    minAvailable: 1
                }

crooked-jelly-50877

10/22/2019, 10:53 PM

I seem to be able to manually run

helm template

without getting rendering errors