pulumi-community #general

general

fast-magician-55948

09/15/2020, 9:34 PM

OK I think I found https://github.com/pulumi/examples/tree/master/gcp-py-serverless-raw that does what I want. I'll see if I can make it work.

red-area-47037

09/15/2020, 10:16 PM

Quick question, is there a way to configure the retry behaviour in Pulumi … e.g. I am currently setting up a GKE environment and from time to time a api service is not available for the first 90s… I would like to configure Pulumi to backoff later retires to ensure that later retries happen after a longer time once the env has stablized

kind-mechanic-53546

09/16/2020, 3:13 AM

is there any way to

await

a resource creation? I have to render a helm chart to a directory via

renderYamlToDirectory

k8s.Provider

and then take the output of that directory filtered via fs.promises.readdir
filtered then passed to k8s.yaml.ConfigGroup
The issue is that my readdir will be processed before the provider has finished rendering I could do it via a custom component but it's more work

chilly-rainbow-79265

09/16/2020, 10:03 AM

Hi all I am trying to have my backend as azure

export AZURE_STORAGE_ACCOUNT=XXXXX
export AZURE_STORAGE_KEY=XXXXX

pulumi new --force
error:
Sorry, could not create stack 'dev': could not create stack: An IO error occurred while writing the new snapshot file: blob (key ".pulumi/stacks/dev.json") (code=NotFound): -> <http://github.com/Azure/azure-storage-blob-go/azblob.newStorageError|github.com/Azure/azure-storage-blob-go/azblob.newStorageError>, /home/travis/gopath/pkg/mod/github.com/!azure/azure-storage-blob-go@v0.9.0/azblob/zc_storage_error.go:42
===== RESPONSE ERROR (ServiceCode=ResourceNotFound) =====
Description=The specified resource does not exist.
RequestId:69f212d5-801e-001a-1910-8c2575000000
Time:2020-09-16T10:02:29.5219649Z, Details:
   Code: ResourceNotFound
   PUT <https://pulumicloud.blob.core.windows.net/pulumistate/.pulumi/stacks/dev.json?timeout=61>
   Content-Length: [70]
   User-Agent: [go-cloud/blob/0.1.0 Azure-Storage/0.9 (go1.14.7; linux)]
   X-Ms-Blob-Cache-Control: []
   X-Ms-Blob-Content-Disposition: []
   X-Ms-Blob-Content-Encoding: []
   X-Ms-Blob-Content-Language: []
   X-Ms-Blob-Content-Md5: [WDyPMum8QM5L4bgdHF8ieA==]
   X-Ms-Blob-Content-Type: [text/plain; charset=utf-8]
   X-Ms-Blob-Type: [BlockBlob]
   X-Ms-Client-Request-Id: [ba87d3af-02c7-45e0-6b1e-550c44d0adcd]
   X-Ms-Version: [2018-11-09]
   --------------------------------------------------------------------------------
   RESPONSE Status: 404 The specified resource does not exist.
   Content-Length: [223]
   Content-Type: [application/xml]
   Date: [Wed, 16 Sep 2020 10:02:29 GMT]
   Server: [Windows-Azure-Blob/1.0 Microsoft-HTTPAPI/2.0]
   X-Ms-Error-Code: [ResourceNotFound]
   X-Ms-Request-Id: [69f212d5-801e-001a-1910-8c2575000000]
   X-Ms-Version: [2018-11-09]

This is a new project all together

incalculable-dream-27508

09/16/2020, 10:17 AM

Ah, and now that i'm trying to use outputs directly in other resources I got one of those errors discussed yesterday that don't really mention where the issue is.

Untitled

incalculable-dream-27508

09/16/2020, 10:21 AM

That's the snippet of code where I'm trying to use the outputs, and that to my understanding produces the above error.

Untitled

chilly-rainbow-79265

09/16/2020, 10:26 AM

@broad-dog-22463 Hi, is there any support from

<docker://pulumi/actions>

for self managed states like blob ? i was trying to do a CD with pulumi and state in blob

incalculable-dream-27508

09/16/2020, 1:41 PM

Huh.

INSTANCES_API = [
    openstack.compute.Instance(
...
        __opts__=pulumi.ResourceOptions(provider=REGIONS[dc]))
    for i in range(COUNT_API) for dc in REGIONS
]

and later

exported_api = {
    srv.name.apply: srv.access_ip_v4.apply
    for srv in INSTANCES_API
}

So far so good, and

pulumi preview

works with this. But, when I add

pulumi.export('nodes_api', exported_api)

suddenly I get this error.

Untitled

incalculable-dream-27508

09/16/2020, 3:59 PM

"WebADM internal traffic: %s %s from %s" % (dc, port, srv.name)

results in names like

WebADM internal traffic: ams rsignd from <pulumi.output.Output object at 0x7eff0ad30100>

so how should I be accessing this? https://www.pulumi.com/docs/intro/concepts/programming-model/#outputs wasn't very helpful regarding that

silly-address-30441

09/16/2020, 5:18 PM

I'm just getting started with Pulumi, so I might not be headed in the right direction at all. Do people generally use different AWS accounts for their prod vs dev stacks? If so, how do you associate a particular AWS account with a particular stack?

faint-dress-91709

09/16/2020, 5:31 PM

hello folks, I’m just getting started with Pulumi. I have experience with Cloudformation, Terraform, and CDK, so some concepts are familiar 🙂 I’m trying to get it working with a AWS profile and am getting an error: “Error: providerCredentialOpts and AWS_PROFILE must be set together”. I tried googling around for this, but no luck so far. I’m sure the error here is between keyboard and monitor, so let me know if I’m doing something stupid ;-)

incalculable-dream-27508

09/16/2020, 5:38 PM

How do people generate unique names for resources when said resources are dynamically generated based on other resources? Namely, because I need to use IPs in security group rules, I need to create a rule per machine created - and apparently can't use machine name, because trying to get that only gets me an ugly (and changing) object reference.

many-psychiatrist-74327

09/16/2020, 6:26 PM

👋 hello! is there a way to call/run pulumi directly from code (that is, instead of running
pulumi
from the terminal)? Context: let’s say I already have some Python code that does a bunch of infrastructure provisioning already, and I want to incorporate some pulumi-managed resources in the same place. The hack I’m doing right now is creating a subprocess that calls

pulumi

, which in turn will end up running some more python code. This seems silly, and I wonder if there’s a way to “cut the middleman”? I know (well, I think) that running

pulumi

stands up a local grpc server, and then runs the Python code, which in turn sends rpcs to the local server to register the resources as it creates them. Could there be a way to run that local server as a daemon maybe, and then just run my Python code normally, without having to create the subprocess mentioned above?

millions-furniture-75402

09/16/2020, 6:34 PM

How do I write an

Output<string>

to a

config.json

tall-scientist-89115

09/16/2020, 6:43 PM

I'm trying to update a cloudfront distriution "Security Policy" from "TLSv1" to "TLSv1.2_2019" with Pulumi (I can do it through the UI no problem but now my state is out of sync), but don't see that property in your distribution docs..the info blurb in the UI for that property says "A security policy determines two settings: the SSL/TLS protocol that CloudFront uses to communicate with viewers, and the cipher that CloudFront uses to encrypt the content that it returns to viewers." How do I set this state with pulumi? (edit: Thank you! I apologize for not putting this in the #aws channel)

ancient-megabyte-79588

09/16/2020, 9:57 PM

Is there a way to run a custom function on

pulumi destroy

... We have

const setupAgent = new azure.compute.Extension(name + "serverSetupAgent",<snip>);

that installs and configures an AzDO Deployment Agent ... If I don't "unregister" the agent with AzDO when I tear the server down, it is left as a dangling registration in AzDO and I can't create another agent with that name. I need to run a de-registration script on

pulumi destroy

or something comparable, but I don't see any where to hook into that.

faint-dress-91709

09/16/2020, 10:09 PM

Hi folks - can I set a launch template on a EKS nodegroup? I don’t see any reference to that in the docs here: https://www.pulumi.com/docs/reference/pkg/nodejs/pulumi/aws/eks/#NodeGroupArgs but it’s available in terraform: https://registry.terraform.io/providers/hashicorp/aws/latest/docs/resources/eks_node_group

chilly-rainbow-79265

09/16/2020, 11:42 PM

HI while running github action using

pulumi/action

for creating azure vm, I'm getting error

ws-ts-azure-comp:webserver:WebServer es-1 create 
    pulumi:providers:azure default  error: no resource plugin 'azure' found in the workspace or on your $PATH

steep-caravan-65104

09/17/2020, 2:48 AM

Hi all. We are using the GCP KMS secrets provider for pulumi as per https://www.pulumi.com/docs/intro/concepts/config/#google-cloud-key-management-service-kms with a separate GCP KMS key used specifically for pulumi key encryption/decryption. We have an issue where users with only encrypt access on this pulumi key are NOT able to encrypt secrets using the command

pulumi config set --secret

They receive the following error:

error: constructing secrets manager of type "cloud": secrets (code=PermissionDenied): rpc error: code = PermissionDenied desc = Permission 'cloudkms.cryptoKeyVersions.useToDecrypt' denied on resource 'projects/<gcp-creds-project>/locations/global/keyRings/global-keyring/cryptoKeys/pulumi-secret' (or it may not exist).

With a GCP user or service account which has encrypt + decrypt permissions on the key, this issue is not seen. Does anyone have ideas about why a DECRYPT key permission would be required to ENCRYPT a secret with pulumi using

gcpkms

? This sounds like a pulumi bug to me, could someone suggest workarounds if possible?

chilly-rainbow-79265

09/17/2020, 5:09 AM

Hi, Is there any way to integrate pulumi with short lived token from cloud provider 1. For instance, If I want to create vm using azure SP which is short lived. I am not sure, what to do here, checking for how you guys do that ?

steep-angle-29984

09/17/2020, 7:27 AM

I'm interested in how other Pulumi users install and configure their VMs as Pulumi doesn't support "configuration management" very well. I haven't found any integration with configuration management tools like Ansible, Chef, or Puppet. So how do you do this?

chilly-rainbow-79265

09/17/2020, 9:14 AM

while doing pulumi preview, it shows all go packes and vendor info. Is there any way to hide that. It'll be pretty much useful in CI/CD commenting

bitter-application-91815

09/17/2020, 9:47 AM

hi there, just looking over how to apply alb but not from typescript but from go, what's

cluster.provider

here

bitter-application-91815

09/17/2020, 9:47 AM

https://www.pulumi.com/blog/kubernetes-ingress-with-aws-alb-ingress-controller-and-pulumi-crosswalk/

incalculable-dream-27508

09/17/2020, 11:57 AM

https://github.com/pulumi/examples/tree/9cb78d2cbcf37a441032a5b6cb4e9f96ff9c8a0a/azure-py-virtual-data-center says

Note that because most resources are auto-named, the trailing dashes that you see above will actually be followed by random suffixes that appear in the Outputs and in Azure.

which would solve the issue I'm fighting with. But I'm reading the code, and I don't see yet how it's achieved?

limited-rainbow-51650

09/17/2020, 3:11 PM

In one stack, I can export an object like structure:

Current stack outputs (2):
    OUTPUT         VALUE
    mysqlAdmin     {"password":"<redacted>","user":"mysqladmin"}
    mysqlEndpoint  <redacted>.<http://database.azure.com|database.azure.com>

But how can I retrieve it as an object via

StackReference

? I seem to miss the call

requireObject<T>(key)

similar like on

Config

. How can I retrieve this info in another stack?

bitter-application-91815

09/17/2020, 3:24 PM

Having an issue applying a helm chart to my newly created kubs cluster,

unknown Invoke type "kubernetes:helm:template"

bitter-application-91815

09/17/2020, 3:25 PM

want to try this one

bitter-application-91815

09/17/2020, 3:25 PM

<https://hub.helm.sh/charts/incubator/aws-alb-ingress-controller>

bitter-application-91815

09/17/2020, 3:25 PM

func (w *worker) applyAlbCharts(ctx *pulumi.Context, namespace string) error {
	_, err := helm.NewChart(ctx, fmt.Sprintf("axiom-alb-charts-%s", w.targetEnv), helm.ChartArgs{
		Chart:     pulumi.String("incubator/aws-alb-ingress-controller"),
		Version:   pulumi.String("1.0.2"),
		Namespace: pulumi.String(namespace),
		FetchArgs: helm.FetchArgs{
			Repo: pulumi.String("<https://kubernetes-charts.storage.googleapis.com/>"),
		},
		Values: pulumi.Map{
			"clusterName":           w.eksCluster.Name,
			"autoDiscoverAwsRegion": pulumi.Bool(true),
			"autoDiscoverAwsVpcID":  pulumi.Bool(true),
		},
	})
	if err != nil {
		return err
	}

	return nil
}

Title

bitter-application-91815

09/17/2020, 3:25 PM

func (w *worker) applyAlbCharts(ctx *pulumi.Context, namespace string) error {
	_, err := helm.NewChart(ctx, fmt.Sprintf("axiom-alb-charts-%s", w.targetEnv), helm.ChartArgs{
		Chart:     pulumi.String("incubator/aws-alb-ingress-controller"),
		Version:   pulumi.String("1.0.2"),
		Namespace: pulumi.String(namespace),
		FetchArgs: helm.FetchArgs{
			Repo: pulumi.String("<https://kubernetes-charts.storage.googleapis.com/>"),
		},
		Values: pulumi.Map{
			"clusterName":           w.eksCluster.Name,
			"autoDiscoverAwsRegion": pulumi.Bool(true),
			"autoDiscoverAwsVpcID":  pulumi.Bool(true),
		},
	})
	if err != nil {
		return err
	}

	return nil
}

View count: 3