pulumi-community #general

Join the conversation Join Slack

general

quiet-wolf-18467

10/02/2020, 9:43 AM

Does anyone know if it's possible to tag as resource in such a way that any change will delete and recreate?

elegant-island-39916

10/02/2020, 10:28 AM

Hi,

k8s.core.v1.Namespace

targets the incorrect k8s cluster (from

~/.kube/config

) even if

opts.providers.kubernetes

is specified e.g.:

ci_namespace = k8s.core.v1.Namespace(
    "ci-svcs",
    opts=ResourceOptions(providers={"kubernetes": eks_cluster.provider}),
)

Yet

pulumi_kubernetes.ConfigFile

does not have this issue and targets the correct cluster. Is there a parameter I am missing on

k8s.core.v1.Namespace

to ensure it targets the providers cluster, rather than my globally configured

~/.kube/config

broad-dog-22463

10/02/2020, 10:43 AM

I believe your ResourceOptions are incorrect @elegant-island-39916

broad-dog-22463

10/02/2020, 10:43 AM

providers isn't a map

broad-dog-22463

10/02/2020, 10:43 AM

elegant-island-39916

10/02/2020, 11:00 AM

@broad-dog-22463 in the definition of

ResourceOptions

providers: Optional[Mapping[str, 'ProviderResource']]

broad-dog-22463

10/02/2020, 11:00 AM

I usually do it as follows

broad-dog-22463

10/02/2020, 11:00 AM

ResourceOptions(provider=k8s_provider)

elegant-island-39916

10/02/2020, 11:00 AM

cool thanks, I'll try that out!

elegant-island-39916

10/02/2020, 11:16 AM

@broad-dog-22463 that worked a treat, thanks a ton! 🙂

broad-dog-22463

10/02/2020, 11:16 AM

nps! sorry that wasn't more clear 😕

little-wolf-72041

10/02/2020, 12:38 PM

Hey does someone know if pulumi previews will get faster if we upgrade to a higher plan?

acoustic-leather-88378

10/02/2020, 1:47 PM

Hi all, I'm trying to wrap https://github.com/mschilz/terraform-provider-infoblox Terraform provider as a Pulumi provider following the steps here: https://github.com/pulumi/pulumi-tf-provider-boilerplate#creating-a-pulumi-terraform-bridge-provider Problem I'm getting on the

make

step is two fold:

make
================
infoblox Package
================
BUILD:
[ -x /usr/local/bin/pulumi ] || curl -fsSL <https://get.pulumi.com> | sh
cd provider && go install -ldflags "-X <http://github.com/pulumi/pulumi-infoblox/provider/pkg/version.Version=v0.0.1-alpha.1593782371+g931a267.dirty|github.com/pulumi/pulumi-infoblox/provider/pkg/version.Version=v0.0.1-alpha.1593782371+g931a267.dirty>" <http://github.com/pulumi/pulumi-infoblox/provider/cmd/pulumi-tfgen-infoblox|github.com/pulumi/pulumi-infoblox/provider/cmd/pulumi-tfgen-infoblox>
# <http://github.com/pulumi/pulumi-infoblox/provider|github.com/pulumi/pulumi-infoblox/provider>
./resources.go:90:26: impossible type assertion:
	*"<http://github.com/hashicorp/terraform-plugin-sdk/helper/schema|github.com/hashicorp/terraform-plugin-sdk/helper/schema>".Provider does not implement "<http://github.com/hashicorp/terraform/terraform|github.com/hashicorp/terraform/terraform>".ResourceProvider (wrong type for Apply method)
		have Apply(*"<http://github.com/hashicorp/terraform-plugin-sdk/terraform|github.com/hashicorp/terraform-plugin-sdk/terraform>".InstanceInfo, *"<http://github.com/hashicorp/terraform-plugin-sdk/terraform|github.com/hashicorp/terraform-plugin-sdk/terraform>".InstanceState, *"<http://github.com/hashicorp/terraform-plugin-sdk/terraform|github.com/hashicorp/terraform-plugin-sdk/terraform>".InstanceDiff) (*"<http://github.com/hashicorp/terraform-plugin-sdk/terraform|github.com/hashicorp/terraform-plugin-sdk/terraform>".InstanceState, error)
		want Apply(*"<http://github.com/hashicorp/terraform/terraform|github.com/hashicorp/terraform/terraform>".InstanceInfo, *"<http://github.com/hashicorp/terraform/terraform|github.com/hashicorp/terraform/terraform>".InstanceState, *"<http://github.com/hashicorp/terraform/terraform|github.com/hashicorp/terraform/terraform>".InstanceDiff) (*"<http://github.com/hashicorp/terraform/terraform|github.com/hashicorp/terraform/terraform>".InstanceState, error)
./resources.go:132:4: unknown field 'AsyncDataSources' in struct literal of type tfbridge.JavaScriptInfo
make: *** [tfgen] Error 2

1. The provider (

p := infoblox.Provider().(*schema.Provider)

) seems to not adhere to the expected interface 2. Missing

AsyncDataSources

JavaScriptInfo

here Any ideas?

hallowed-beach-15050

10/02/2020, 4:18 PM

https://twitter.com/adamhjk/status/1312051814049542145?s=20

😛-master: 1

🎯 1

😛artypus-8bit: 1

gentle-account-13294

10/02/2020, 5:38 PM

Hi Community, We are going all in on Pulumi but would like some recommendations on how you all are handling

development

infrastructure workflows . I.e. while testing changes to infrastructure ,what workflows are you used to... (options in thread)

busy-umbrella-36067

10/02/2020, 5:48 PM

Is there any way to disable the encryption salt when using S3 state?

kind-school-28825

10/03/2020, 9:44 PM

Are there any plans to open up a community forum for Pulumi on the official website to ease of looking up previously reported issues or questions from other users? Imo slack is a bit painful when it comes to this

rough-parrot-6341

10/04/2020, 8:02 PM

So I've been experimenting with deploying kubernetes clusters and resources on said cluster with Pulumi, and hit an issue that I'm not sure what would be the proper way to handle correctly?

rough-parrot-6341

10/04/2020, 8:03 PM

Simply put, attempting to deploy a helm chart with pulumi that fails to create some resources (due to the healthcheck failing, not because the resources aren't actually created) makes Pulumi think the resources don't exist when they do

rough-parrot-6341

10/04/2020, 8:03 PM

and bam you already have configuration drift

rough-parrot-6341

10/04/2020, 8:03 PM

an example trying to deploy a helm chart that refused to deploy properly on the DO kubernetes

rough-parrot-6341

10/04/2020, 8:04 PM

now Pulumi thinks these 3 failed resources don't exist on the cluster while they do, but just have failing healthchecks

rough-parrot-6341

10/04/2020, 8:04 PM

e.g.

rough-parrot-6341

10/04/2020, 8:07 PM

my understanding from the docs was that Pulumi somehow tries to ensure consistency between resources managed by Pulumi and what it has tracked in it's state, but this doesn't seem to be the case? I can see that the failing resource has a

managed-by: pulumi

label, yet Pulumi didn't clean it up when there should be no resources deployed on the cluster anymore by Pulumi

rough-parrot-6341

10/04/2020, 8:08 PM

so I'm wondering what are you supposed to do in this kind of a scenario. Getting to a state of configuration drift (accidentally at that) was surprisingly easy just by using Pulumi, which makes me a bit afraid to use it for production purposes

rough-parrot-6341

10/04/2020, 8:09 PM

understandably Kubernetes is the kind of resource Pulumi can't wish to completely keep track of, but I'd expect it to properly track the resources it itself deploys (and labels even)

rough-parrot-6341

10/04/2020, 8:11 PM

the actual issue (and why I noticed this) is that if I now try to re-deploy the chart, it will fail due to name conflicts (as Pulumi didn't clean up properly after the previous installation was removed)

rough-parrot-6341

10/04/2020, 8:12 PM

on the other hand, had I used helm for this operation, I'm fairly certain it would have cleaned up after itself correctly

rough-parrot-6341

10/04/2020, 8:48 PM

also how are cluster version upgrades supposed to be handled? is there a way to say

use version x for creating the cluster if it doesn't exist yet, otherwise ignore version difference

rough-parrot-6341

10/04/2020, 8:48 PM

I tried using "latest" on the cluster version as I saw in some guide, but that wanted to replace the cluster every time I ran

pulumi up