Is there any existing module or example creating security groups from some simple yaml or json configuration?

hey all, is anyone using pulumi and azure devops release pipelines together? I have some questions if so and you want to help a man out! 😃

When trying to run

pulumi up

in a typescript project utilizing

@pulumi/pulumi-kubernetes

, the preview shows that all Kubernetes resources would be replaced, since the

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

label ought to be removed. The details show changes along the line of:

[provider: urn:pulumi:dev::project::eks:index:Cluster$pulumi:providers:kubernetes::project-dev-provider::617c0adc-0f4b-47cc-8e2a-319f67902782 => urn:pulumi:dev::project::eks:index:Cluster$pulumi:providers:kubernetes::project-dev-provider::output<string>]
~ metadata  : {
    - labels   : {
        - <http://app.kubernetes.io/managed-by|app.kubernetes.io/managed-by>: "pulumi"
      }
  }

In another project with the same package versions (

0.25.2

) this doesn't happen, however. Has anyone seen a similar behaviour or has an idea what might be causing this?

https://www.pulumi.com/blog/ci/cd-made-easy-with-pulumi-and-azure-pipelines did this article got moved somewhere? was loading a few days ago and the content was very relevant for a project I’ve resumed

After updating to

@pulumi/kubernetes

version 0.25.2 I’m seeing this error:

warning: resource plugin kubernetes is expected to have version >=0.25.2, but has 0.25.1-dev.1561392430+g1853784.dirty; the wrong version may be on your path, or this may be a bug in the plugin

I tried removing all plugins with

pulumi plugin rm --all

and then reinstalling with

npm i

to refresh it but I still get the warning. Is the package set up to install the correct plugin?

@white-balloon-205 How does this compare to Pulumi? Have you guys seen this yet? thoughts? https://docs.aws.amazon.com/cdk/latest/guide/home.html

hello. is someone able to point me in the direction of an example for creating a mapping from a domain name (e.g.

<http://example.com|example.com>

) to a service (like the example on: `https://www.pulumi.com/docs/reference/tutorials/aws/tutorial-service`)?

I keep getting

Cannot read property name of undefined

with

envFrom: configMaps.map(function(configMap) {
              return {
                configMapRef: {
                  name: configMap.metadata.name
                }
              };
            }),

where

configMaps

is an array of resources.

Anyone have any ideas?

Ah, figured it out

Bad design of my component resource

I have an instance where pulumi tries to "replace" a resource after an upgrade of pulumi-gcp from 0.17.1 to 0.18.12 even though nothing in that resource has changed. This problem is similar to https://github.com/pulumi/pulumi-gcp/issues/88 however in this case it's a VM that pulumi attempts to replace.

+-gcp:compute/instance:Instance: (replace)
        [id=vpn-00a]
        [urn=urn:pulumi:gcp-project-solvvy-prod::gcp-project-solvvy-prod::gcp:compute/instance:Instance::instance]
      ~ networkInterfaces: [
          ~ [0]: {
                  + __defaults   : []
                  ~ accessConfigs: [
                      ~ [0]: {
                              + __defaults: []
                                natIp     : "32.247.25.46"
                            }
                    ]
                    networkIp    : "10.0.16.10"
                    subnetwork   : "us-west1/vpn-subnetwork"
                }
        ]

there's absolutely 0 change on that resource except that i upgraded the pulumi gcp version do you have a quick idea how to resolve this? I'm fine editing the state if that helps, but I need to know what the state should look like for pulumi to not trigger a replace.

here's the source code for the resource:

const vpnInstance = new gcp.compute.Instance(
    'instance',
    {
        name: 'vpn-00a',
        zone: 'us-west2-b',
        machineType: 'n1-standard-1',
        canIpForward: true,
        deletionProtection: true,
        metadata: { 
            'startup-script-url': startupScriptURL
        },
        scheduling: {
            automaticRestart: true
        },
        bootDisk: {
            initializeParams: {
                image: 'ubuntu-os-cloud/ubuntu-minimal-1804-lts',
            },
        },
        networkInterfaces: [{
            subnetwork: vpnSubNetwork.id,
            networkIp: vpnInternalStaticIP,
            accessConfigs: [{
                natIp: vpnExternalIP.address
            }]
        }],
        serviceAccount: {
            scopes: ['<https://www.googleapis.com/auth/cloud-platform>'],
        },
    },
    { dependsOn: [vpnFirewall] });

Weird. Pulumi refresh is detecting the changes, but when I run pulumi up, it does not try to apply the changes. There is currently 1 replica, I refreshed (because it should have been 2 but it is 1 for some reason), ran up, and it shows no changes.

b

typescript question... having trouble compiling some typescript that required me to add a

global.d.ts

declaration file. TSC in vscode understands the declartion is there but

pulumi

doesn't seem to

is there a trick to configuring pulumi's special tsconfig to help it find declarations

?

It seems that Pulumi wants to replace all k8s resources because the provider of the underlying EKS cluster is marked to be replaced as well. The EKS cluster was created via

new eks.Cluster(stackName, {
  instanceType: "t2.medium",
  desiredCapacity: 2,
  minSize: 2,
  maxSize: 4,
  nodeRootVolumeSize: 60
});

and that code hasn't changed in months. Does anybody have an idea what might be going on and how to fix this? The preview just states that the provider will be replaced but no diff is provided.

Does Pulumi support using Ansible as configuration management when deploying an EC2 instance? How would I provision using Type Script?

Anyone know how to grant

roles/errorreporting.viewer

? Does this have to be done as a project iammember policy?

Howdy, I'm testing https://github.com/pulumi/pulumi/pull/2893 the new import functionality in Pulumi a lot and have found some odd behavior I think? I didn't want to keep spamming the PR, and thought here would be a good place to go through it. I'm importing an ELB in Python, and I've only set these parameters to the ELB

elb = elb.LoadBalancer(resource_name=elb_name, opts=resource_opts, internal=scheme, name=elb_name, listeners=listeners, tags=tags)

The ELB I have configured in AWS explicitly has configuration I have not set here ^ however, Pulumi preview/up think that my state defined in Python 100% matches the state from AWS. I'm missing ELBHealthchecks, security groups, AZ's, and a bunch of other options. I'm surprised Pulumi didn't yell at me that I was missing these settings?

You can see that Pulumi preview does show that Pulumi is aware that the missing options do exist and are set in AWS, but doesn't complain that my code is missing those options. Pulumi goes through with the import successfully, but I think this is bad behavior since I'm missing random options I had configured.

I can't quite figure out how Pulumi is deciding what properties must be defined before import, and which seem to just be optional. If I leave off tags, or listeners, Pulumi yells at me that there's a config difference, but if I leave out subnets, sourceSecurityGroup, securityGroups, idleTimeout, healthCheck, crossZoneLoadBalancing, connectionDraining, connectionDrainingTimeout, or availabilityZones, Pulumi doesn't care.

Ultimately, whether or not there is a difference between the actual and desired state is up to the resource provider. In the case of this particular resource, my guess is that the underlying AWS TF provider treats the properties that are set but not deemed required as potential output properties, and therefore ignores any difference if the property is unset in the resource's inputs.

I agree that this behavior is a little weird. If you set the options that aren't deemed necessary, does your import still succeed?

Yes, it does still succeed, but I just suddenly have floating configs that aren't tracked in my code anywhere which is unfortunate.

I was looking at the repo layout of

pulumi-awsx

and

pulumi-eks

to see what could be reused for my own pulumi modules/packages. I found out the build setup is based on Makefiles. With Bazel Build (https://bazel.build/) these days, it’s fairly easy to create re-usable build rules which means that replicating a build setup shouldn’t have Makefiles duplicated. Any specific reason for this old-fashioned (😄) build tooling?

anyone have experience with pulumi github actions?

How do you access the resources in a helm chart? Trying to get the name of a resource.