@ambitious-ram-5811 the new “Pulumi watch” stuff is great for iterative development too!

I am very confused, I am trying to put a deployment in a special namespace I created. But the Yaml Pulumi generates points to "default" even though I specify a different namespace.

morning all. ive got a k8s cluster being provisioned and it all works very nicely - i'm using pulumi in typescript and using the digitalocean provider. the only problem ive got is with regards to refreshing the kubeconfig. when i first build the cluster, it generates a kubeconfig which i then set to a k8s provider so i can use it internally. the problem is that the kubeconfig that DO creates expires after 7 days. and because pulumi doesnt know this, it never calls DO to regenerate the token. this is what im doing so far:

const k8s = new digitalocean.KubernetesCluster('Configure k8s', stackConfig.k8s);

const k8sConfig = k8s.kubeConfigs[0].rawConfig;

const k8sProvider = new kubernetes.Provider('Get k8s provider', {
  kubeconfig: k8sConfig,
});

the

k8sProvider

then is set to

provider

options any time i make an additional k8s call, eg

const helmServiceAccount = new kubernetes.core.v1.ServiceAccount(
  'Helm RBAC service account',
  {
    apiVersion: 'v1',
    kind: 'ServiceAccount',
    metadata: {
      name: 'tiller',
      namespace: 'kube-system',
    },
  },
  {
    provider: k8sProvider,
  },
);

as i said, this all works fine for the first 7 days then breaks because the kubeconfig has expired within digitalocean

aaaah, ive found it. it's

pulumi update

- id been playing around with

pulumi refresh

but doesnt work for <reasons> https://github.com/pulumi/pulumi/issues/2710

I’ve tried using

apply

but can’t figure out what to do with it.

What’s the correct way to use

requireSecret

as an argument to a cloud spec? I see that in https://github.com/pulumi/examples/blob/fdf37ddecedb10dcdfa77362f8ece49c52011caa/aws-stackreference-architecture/database/src/index.ts#L30 nothing special is going on but when I follow that pattern like:

const config = new Config();
const dbPassword = config.requireSecret('dbPassword');

// stuff....

const postgresUser = new gcp.sql.User('master', {
  name: dbUser,
  password: dbPassword, // this line errors out. I've also tried dbPassword.get()
  instance: postgresDB.name,
  project: project
});

I get the following error:

index.ts(187,3): error TS2322: Type 'Output<string>' is not assignable to type 'string | Promise<string> | OutputInstance<string> | undefined'.
      Type 'Output<string>' is not assignable to type 'OutputInstance<string>'.
        Types of property 'apply' are incompatible.
          Type '{ <U>(func: (t: string) => Promise<U>): import("/Users/paymahn/qwil/[secret]/pulumi/infra/node_modules/@pulumi/pulumi/output").Output<U>; <U>(func: (t: string) => import("/Users/paymahn/qwil/[secret]/pulumi/infra/node_modules/@pulumi/pulumi/output").OutputInstance<U>): import("/Users/paymahn/qwil/[secret]/pulumi/infra/nod...' is not assignable to type '{ <U>(func: (t: string) => Promise<U>): import("/Users/paymahn/qwil/[secret]/pulumi/node_modules/@pulumi/pulumi/output").Output<U>; <U>(func: (t: string) => import("/Users/paymahn/qwil/[secret]/pulumi/node_modules/@pulumi/pulumi/output").OutputInstance<U>): import("/Users/paymahn/qwil/[secret]/pulumi/node_modules/@pulumi/...'.
            Type 'import("/Users/paymahn/qwil/[secret]/pulumi/infra/node_modules/@pulumi/pulumi/output").Output<any>' is not assignable to type 'import("/Users/paymahn/qwil/[secret]/pulumi/node_modules/@pulumi/pulumi/output").Output<any>'.
              Type 'OutputInstance<any>' is not assignable to type 'Output<any>'.
                Type 'OutputInstance<any>' is not assignable to type 'OutputInstance<any> & LiftedObject<any, string>'.
                  Type 'import("/Users/paymahn/qwil/[secret]/pulumi/infra/node_modules/@pulumi/pulumi/output").OutputInstance<any>' is not assignable to type 'import("/Users/paymahn/qwil/[secret]/pulumi/node_modules/@pulumi/pulumi/output").OutputInstance<any>'.
                    Types of property 'apply' are incompatible.
                      Type '{ <U>(func: (t: any) => Promise<U>): import("/Users/paymahn/qwil/[secret]/pulumi/infra/node_modules/@pulumi/pulumi/output").Output<U>; <U>(func: (t: any) => import("/Users/paymahn/qwil/[secret]/pulumi/infra/node_modules/@pulumi/pulumi/output").OutputInstance<U>): import("/Users/paymahn/qwil/[secret]/pulumi/infra/node_modu...' is not assignable to type '{ <U>(func: (t: any) => Promise<U>): import("/Users/paymahn/qwil/[secret]/pulumi/node_modules/@pulumi/pulumi/output").Output<U>; <U>(func: (t: any) => import("/Users/paymahn/qwil/[secret]/pulumi/node_modules/@pulumi/pulumi/output").OutputInstance<U>): import("/Users/paymahn/qwil/[secret]/pulumi/node_modules/@pulumi/pulumi...'.
                        Type 'import("/Users/paymahn/qwil/[secret]/pulumi/infra/node_modules/@pulumi/pulumi/output").Output<any>' is not assignable to type 'import("/Users/paymahn/qwil/[secret]/pulumi/node_modules/@pulumi/pulumi/output").Output<any>'.
                          Type 'OutputInstance<any>' is not assignable to type 'Output<any>'.
                            Type 'OutputInstance<any>' is not assignable to type 'OutputInstance<any> & LiftedObject<any, string>'.
                              Type 'OutputInstance<any>' is not assignable to type 'LiftedObject<any, string>'.
                                Index signature is missing in type 'OutputInstance<any>'.

I’ve tried calling

.get()

on

dbPassword

but that results in another error of

Error: Cannot call '.get' during update or preview.
    To manipulate the value of this Output, use '.apply' instead.

I’ve declared a docker image like:

const appImage = new docker.Image("ledgerImage", {
    imageName: '<http://us.gcr.io/qwil-build/ledger|us.gcr.io/qwil-build/ledger>',
    build: "../../",
});

When I run

pulumi preview

and look at the output in app.pulumi.com I see the line

Successfully pushed to docker

in the Diagnostic Log view. Where did pulumi push this image? When I visit my companies gcr registry I don’t see any new images - the latest image that was uploaded is from 2 days ago.

Can we run aws cli commands inside pulumi?

Did anybody already use Azure AKS (Kubernetes service) with multiple node pools (aka VMSS)? I’m specifically looking for some hints on how to handle the different pools together with custom networking (vnet, subnet), azure-cni and the Standard SKU load balancer required for the vm scale sets. Thanks for any pointer!

any complex examples with typescript out there? Like classes, interfaces and the like? Just need some advice on how to set it up in an enterprise environment. going beyond the simple index.ts examples ;)

How does Pulumi handle Kubernetes API deprecations these days, e.g.: https://kubernetes.io/blog/2019/07/18/api-deprecations-in-1-16/ We have some

apiVersion: extensions/v1beta1

resources that we want to change to, e.g.

apiVersion: <http://networking.k8s.io/v1|networking.k8s.io/v1>

. My experience with Helm is that it kinda fucks up here, because for e.g.: a deployment named

foo

, we can't tell Helm "don't change anything with this resource, just update Tiller with what the new API version is because k8s allows us to access a deployment via both extensions/v1beta1 and apps/v1". The result is we have to rename everything to

foo-v1

, experiencing some downtime because Helm will tear down the old deployment, stand up the new deployment, and often other stuff is also infected with the name via Helm templating so the service, the ingress, etc all get replaced and/or destroyed and recreated as the new name. Is there a way to make Pulumi just... accept that these resources are now

<http://networking.k8s.io/v1|networking.k8s.io/v1>

or

apps/v1

? That is, no update/replace/etc needed.

Has anyone here seen

AssertionError: Failure: Decoder hit an error

error when trying to

pulumi up

?

I deleted my k8s cluster from the AWS console. Now my Pulumi stack fails with "unable to load schema information from the API server". Is there a way to simply tell Pulumi that the EKS cluster is gone so all of the k8s resources are also gone?

I'm trying to use pulumi with the program in one directory and the stack files in another. e.g.

❯ tree
.
├── program
│   └── index.ts
└── stack
    ├── Pulumi.my-stack.yaml
    └── Pulumi.yaml

is this feasible?

Hello all, Quick question regarding to Azure SQl Server. How could I set "Allow Azure services and resources to access this server" from pulumi when provision the server and how to restrict to particular AzureDevOps instance?

Hrm, I recently added a FargateTaskDefinition into my pulumi project... it works great, but it means now that every time I

pulumi up

it builds the image and pushes it to the container registry, even if there were no changes. This takes a fair bit of time, and requires that I have Docker running even if I'm not actively doing any Docker work. Anyone know of a way to have it only do the docker work if changes were actually made to the docker files?

Hi All, How could we specify the vCore of SQL Db when creating new database with General Purpose edition? Thanks all 🙂

Is it possible to install the Pulumi CLI via NPM? Maybe something like

yarn add @pulumi/cli

?

@white-balloon-205 @future-barista-68134 How about creating extra Slack channels for major Pulumi providers? Then questions specific to a provider can be easier filtered by joining only the channels you know of. Not distinguished by the programming language used. I think that doesn’t matter much. In case, the question can be cross-referenced to one of the language channels, though.

Are we able to use an Azure Government blob storage account as the backend to store states? I have set my storage account and key environment variables properly, and tried using both the

pulumi login azblob://<container>

method as well as

pulumi login -c <account>.<http://blob.core.usgovcloudapi.net|blob.core.usgovcloudapi.net>

. However, azblob://<container> always defaults to use

<account>.<http://blob.core.windows.net|blob.core.windows.net>

, and whenever I do "pulumi login -c", I am asked for an access token, and any access key or sas token I use results in authorization errors. Am I doing something wrong, or is this just not possible at the moment?

hey guys I’m getting this error on one of my resources when ever I run a pulumi up

error: unable to load schema information from the API server: Get <https://35.227.104.48/openapi/v2?timeout=32s>: dial tcp 35.227.104.48:443: i/o timeout

There can be only one Jay

@ambitious-helmet-24352 @little-tailor-59400 wth. How are we sharing a username?

For whatever reason, I thought slack cared about duplicates

Haha there's nothing wrong with a replica

Yeah it never complained when i created the username funny enough

Weird

Good to know :)

Is there any documentation I can read about

default

values in the

Pulumi.yaml

? They seem to be not working but I imagine I'm doing it wrong.

Hey everyone, I'm facing a weird issue were I'm consistently getting a duplicate resource error for any resource I try to create e.g

const vpcName = `unique-vpc-2334`;
  const vpcArgs = {
    cidrBlock: projectCidrBlock,
    enableDnsHostnames: true,
    enableDnsSupport: true,
    tags: {
      Name: vpcName,
    },
  };
  const vpc = new aws.ec2.Vpc(
    vpcName,
    vpcArgs,
    providerOpts,
  );

It doesn't matter what I change the name to I always end up getting the same error.

error: Duplicate resource URN 'urn:pulumi:ned-pulumi::ned-pulumi::aws:ec2/vpc:Vpc::unique-vpc-2334'; try giving it a unique name

I've also tried changing the provider, deleting and recreating the stack with a new name. I have a feeling that it may be caching values/running twice with the same values. Let me know if anyone has resolved this issue, thanks