but is it even possible to ignore changes to children? the other use cases seem to ignore changes to properties, which seem to be treated differently

ah right, so the issue is that

ignoreChanges

only works for properties defined in the code, not for what pulumi checks for state-sync

the more I try to deploy helm charts with pulumi the more it feels like I shouldn't do that

it just so happens many stacks are distributed as helm charts

If I create my own subclass of

ComponentResource

, is it obliged to create my resources in the constructor? Or can I delay the creation of (some of) them?

Also inside of its constructor, you will allocate any child resources, passing the

parent

option as appropriate to ensure component children are parented correctly.

https://www.pulumi.com/docs/intro/concepts/programming-model/#authoring-a-new-component

Hey folks, wondering if there's a way of unexporting/removing a stack output for good. Couldn't find one in the docs / open issues

is it possible to inject additional config values at runtime outside of tests? I have a secret stored on AWS that I need to inject while using the github provider, I wanted to avoid storing as a config secret and as env var. What I am trying to do is fetching it at runtime for further use but it is not working

Hi, are you planning to add Gtihub Org level secret support?

Greetings from Moldova!

Is it possible to move an entire Pulumi program from one Git repo to another without disrupting any of the resources?

I don't think the program is linked to the git repo, but to the stack

So I'd imagine you're fine as long as you connect it to the correct stack

I don't know how or where does Pulumi retain the project -> stack connection on disk

Hi, I am creating multiple Pulumi projects and wish to share outputs via stack references between them (using an S3 backend). Different projects with the same stack name is not possible so I am resorting to namespaced stack names for each project e.g.

infra-dev

,

services-dev

,

infra-prod

,

services-prod

Though that is a bit of a pain to manage and would make sense to simply namespace on a project level rather than purely via stacks. Is there a way around this that allows setting up multiple projects with the same stack name, similar to how the documentation outlines? https://www.pulumi.com/docs/intro/concepts/organizing-stacks-projects/#inter-stack-dependencies

so I have a function which can return a string (if what I look for is an env var) or an

Output

which will grab that from another place, is there a way to "cast" a string to pulumi output so that the output will be consistent?

> pulumi update --skip-preview --yes -v=6

Updating (catmeme/pulumi-v1-102)
View Live: <https://app.pulumi.com/>...
pulumi:pulumi:Stack gameserver-pulumi-v1-102 creating
@ Updating........

+ pulumi:pulumi:Stack gameserver-pulumi-v1-102 creating error: an unhandled error occurred: Program exited with non-zero exit code: -1

+ pulumi:pulumi:Stack gameserver-pulumi-v1-102 **creating failed** 1 error

Diagnostics:

pulumi:pulumi:Stack (gameserver-pulumi-v1-102):

error: an unhandled error occurred: Program exited with non-zero exit code: -1

Resources:

+ 1 created

Duration: 1m55s

How can I debug this failure further? I’m using -v=6 and not getting any helpful output

try v=9

did not make a difference :-/

Anyway to see why a stack is stuck

Another day, another question about string interpolation 🙂 Is there any clean way to apply a resource into a getResource, e.g.

aws.iam.getPolicyDocument

? The below works, but is nasty, casting to any just to avoid compile time errors about transforming from output to string

const assumeDevelopersRolePolicyData = await aws.iam.getPolicyDocument({
    version: "2012-10-17",
    statements: [
      {
        effect: "Allow",
        actions: ["sts:AssumeRole"],
        resources: [
          <any>(
            pulumi.interpolate`arn:aws:iam::${dev_account_id}:role/${devRole.name}`
          ),
        ],
      },
    ],
  });

Or is the only reason this is working currently is that I'm lucky and it'll fail at some point in the future 🙂

Hi. What is

aws_vpc

that is shown in some of the Python examples? For example: https://www.pulumi.com/docs/reference/pkg/aws/ec2/securitygroup/#example-usage

Is there a way to use the name of a provider in the logical name of a resource if you only have a reference to the provider and not the configuration of the provider? For example:

function createVpc(provider: ProviderResource) {
    new aws.ec2.Vpc(`${provider.id}-main`, {
        cidrBlock: "10.0.0.0/16",
    }, {provider: provider});
}

Using the code exactly as it is in my example yields the error:

error: Duplicate resource URN 'urn:pulumi:main::acddevops-eks::aws:ec2/vpc:Vpc::Calling [toString] on an [Output<T>] is not supported.

Why do I keep seeing:

Outputs:
  - bucket_name: "tlayen-pulumi-demo"

In the output when I have long deleted that output?

I'm not sure but... https://www.pulumi.com/docs/reference/pkg/gcp/firestore/ If I click on "Index" it gives 404. I assume this is a bug

Hi! i'm new using pulumi and I have a doubt .... what is the difference between id and name in resource AWS Nodegroup?

With Pulumi/GCP, what is the equivalent of the

allow all

firewall rule? I'd like to enable all traffic on all ports within a subnetwork

Any idea why today when I do

config.require('role')

that it doesn't expect it to be prefixed with the name as in

foo:role

?

It worked fine yesterday

Oh never mind. It was something else. I have to be honest the error messages in Pulumi are not very good. I thought terraform was bad but this is worse. I don't mean any offence, just some critical feedback.

Long shot, but is anyone using the Okta provider? I can't find a way to add an OAuth app to the Okta dashboard. Thank you.