pulumi-community #general

general

busy-umbrella-36067

10/02/2019, 4:37 PM

Interesting, seems that

pulumirpc.ResourceMonitor/RegisterResource

takes up 85% of the preview time

cool-egg-852

10/02/2019, 4:39 PM

Any reason

<http://pulumi.com/timeoutSeconds|pulumi.com/timeoutSeconds>

is an annotation instead of an option to the pulumi resource? Seems weird to be modifying the resource in order to change the behavior of pulumi. Same goes for

skipAwait

cool-egg-852

10/02/2019, 5:02 PM

Is there any information on when the kubernetes provider 1.2 version will be released? Asking because we’re running into an issue with the hack done for waiting on jobs, and it would be cool to not need to fix the hack further.

acceptable-army-69872

10/02/2019, 5:48 PM

I'm trying to do an import on an s3 bucket, and it wants to change the name of the resource, which causes a warning.

aws:s3:Bucket (bucketName):
  warning: inputs to import do not match the existing resource; importing this resource will fail

This is the change: 
~ bucket: "bucketName" => "bucketName-8c54a03"

code

const config = new pulumi.Config();
const sb: string = config.require("sourceBucket")

const sourceBucket = new aws.s3.Bucket(sb,
                                      {},
                                      { import: sb });

incalculable-engineer-92975

10/02/2019, 6:32 PM

This is a support request for Pulumi. I have a stack that deploys the helm chart 'aws-alb-ingress-controller', then creates an Ingress k8s resource. The combination of those two items results in the creation of an AWS ALB (the controller monitors Ingress resource CRUD and mirrors those changes in an ALB). The works perfectly when doing a 'pulumi up'. Here's the problem point. when doing a 'pulumi destroy', the controller is either destroyed before the Ingress deletion is noticed or the Ingress is deleted after the controller. How do I prevent this from happening? Is it possible to add a delay or hook or anything that will alter this processing? Thanks in advance.

cool-egg-852

10/02/2019, 7:57 PM

I’m curious what people would recommend here: We have a project (CI). We have 4 environments (dev, staging, prod, support). CI runs in support. We need to create a GCP service account in the support stack (we have one stack per env), but then reuse that stack in the other 3 envs so that we can grant it a role in GCP. Given that TypeScript doesn’t allow exporting inside of a conditional, what’s the best way to reference that service account? The issue is due to TS and the way Pulumi requires things to be `export`ed for it to be used in a stack reference.

adamant-dress-73325

10/02/2019, 9:06 PM

@lemon-spoon-91807 Do you know if the stack ref OutputSync bug was fixed? We are doing this in a lot of places, wondering if it was still needed?

const vpcPrivateSubnetIds = (stackRef.requireOutputSync("vpcPrivateSubnetIds") as unknown) as Array<
  pulumi.Output<string>
  >;

miniature-chef-87843

10/03/2019, 1:33 AM

If I wanted to pre-process a function (say, webpack or babel), would I just do a normal webpack build process and

require

the output in a

CallbackFunction

prehistoric-analyst-52684

10/03/2019, 11:07 AM

I cannot get aliases to work to migrate my resources when I move them inside component resources. I have a bunch of resources that didn't have an explicit parent. I created a component resource and moved all existing resource inside. Then I exported the stack and took the `urn`s to use in the

aliases

argument of each existing resource. When I

pulumi up

, all resources are to be recreated. I am pretty sure the urn is correct, as I am taking it from the exported stack file. Here is an example resource:

const logGroup = new aws.cloudwatch.LogGroup(`${name}_cloudwatchLogGroup`, {}, { provider, parent: this, aliases: ["urn:pulumi:edith-lookup-feat::edith::aws:cloudwatch/logGroup:LogGroup::cloudwatchLogGroup_edith-lookup-service"] });

What am I doing wrong?

cool-egg-852

10/03/2019, 12:27 PM

When running pulumi in a dockerized CI environment like CodeFresh, we’re getting:

Previewing update (LinioIT/support):                                                                                              
[resource plugin aws-1.2.1] installing                                                                                            
Downloading plugin: 60.87 MiB / 60.87 MiB  100.00% 4sm1s                                                                          
Moving plugin... done.                                                                                                            
                                                                                                                                  
    pulumi:pulumi:Stack rsin-provider-support running                                                                             
    pulumi:providers:aws default_1_4_0  error: no resource plugin 'aws-v1.4.0' found in the workspace or on your $PATH, install th
e plugin using `pulumi plugin install resource aws v1.4.0`                                                                        
    pulumi:pulumi:Stack rsin-provider-support                                                                                     
    pulumi:providers:aws default_1_4_0  1 error                                                                                   
                                                                                                                                  
Diagnostics:                                                                                                                      
  pulumi:providers:aws (default_1_4_0):                                                                                           
    error: no resource plugin 'aws-v1.4.0' found in the workspace or on your $PATH, install the plugin using `pulumi plugin instal
l resource aws v1.4.0`

Any idea? It seems like pulumi is installing the wrong version, but it is detecting it needs to install the plugin at minimum.

incalculable-engineer-92975

10/03/2019, 4:11 PM

Is there any way to tell Pulumi that a resource was created w/o creating that resource? The Pulumi script creates an ALB as a side-effect and it doesn't get deleted when doing a

pulumi destroy

. I've done dependencies, I've tried delays, nothing works. So I'm looking at this as my last resort.

nice-airport-15607

10/03/2019, 5:16 PM

Hey Pulumers! Is there a way to see the git details that you’re logged into via the cli? When logging into Pulumi console, I log in with my github creds, that being said, I see the latest branch hash in the console, I was curious if there’s a way to see that from my local machine via

pulumi stack...

commands.

busy-umbrella-36067

10/03/2019, 6:07 PM

I’m getting a failure on

pulumi up

with this error:

Assertion failed: (handle->type == UV_TCP || handle->type == UV_TTY || handle->type == UV_NAMED_PIPE), function uv___stream_fd, file ../deps/uv/src/unix/stream.c, line 1640.

tall-terabyte-42834

10/03/2019, 8:48 PM

Hi Pulumi, I'm trying to do some refactoring and asserting an empty

pulumi preview

before making changes. But even before any refactoring, it's already showing that a bunch of stuff will change. Like it would blow away and recreate stuff (which would be bad). What could cause this and how to work past it?

stale-autumn-24797

10/03/2019, 9:07 PM

A question about Pulumi AKS support: Am I able to configure a Kubernetes cluster to use virtual nodes in Pulumi, as described in this article: https://docs.microsoft.com/en-us/azure/aks/virtual-nodes-cli?

billowy-laptop-45963

10/03/2019, 9:47 PM

I'd like to run integration tests outside my current working environment but it seems it's not possible unless I sandbox the process in a container/vm or something due to the way pulumi login is global. Is there a way around this?

early-musician-41645

10/03/2019, 10:04 PM

I know this has come up a lot, but I'm just not seeing the problem:

const zone = aws.route53.Zone.get("external-dns", ingress.externalDnsHostedZoneId);
      let domain: string = zone.name.apply(z => z.replace(/\.$/, ''));
      let zoneId: string = zone.zoneId.apply(z => z);

leads to:

error TS2322: Type 'Output<string>' is not assignable to type 'string'.

Any tips? How do I cast

Output<string>

string

cool-egg-852

10/03/2019, 10:09 PM

Any advances coming up to reduce the amount of time it takes run update a stack with around 50-150 resources?

cool-egg-852

10/03/2019, 10:10 PM

I feel like updating a stack dedicated to istio takes forever (I think there’s around 120 resources in mine)

billowy-laptop-45963

10/03/2019, 10:56 PM

when I do pulumi up it just gets stuck in previewing, I tried -v=3, --skip-preview (it gets stuck in updating), --tracing. Any other way to debug this? Integration tests are stuck in a similar way.

early-intern-90238

10/03/2019, 11:08 PM

Hey guys, quick opinionated question. If your using Kubernetes do you prefer a seperate dev environment, I am currently managing dev with labels but I've had some scenarious where this has been pretty problematic. So I am thinking of using an entirely seperate cluster for dev, stage, and only share canary with production. As for developer machines, is anyone using tilt?

big-nail-28315

10/04/2019, 2:02 PM

I need to create API Gateway VPC Links using Network Load Balancers that were created during provisioning of an Elastic Beanstalk environment. I see that I can get the load balancer id's from the Elastic Beanstalk environment, but is there a way to get the ARN's of those load balancer id's so I can pass them to a

new VPCLink()

? I was thinking that I could import the Beanstalk NLB's, but it doesn't look like the import syntax works on those.

cool-egg-852

10/04/2019, 9:14 PM

Diagnostics:
  pulumi:pulumi:Stack (cert-manager-development):
    error: Error: Command failed: helm fetch jetstack/cert-manager --untar --version v0.10.1 --destination /var/folders/dn/t652qw8j7g5378d66qqng_dh0000gn/T/tmp-10896ZtroXrGxEMe2 --repo https\://charts.jetstack.io
    Error: chart "jetstack/cert-manager" version "v0.10.1" not found in <https://charts.jetstack.io> repository

    Error: chart "jetstack/cert-manager" version "v0.10.1" not found in <https://charts.jetstack.io> repository
    Error: Error: Command failed: helm fetch jetstack/cert-manager --untar --version v0.10.1 --destination /var/folders/dn/t652qw8j7g5378d66qqng_dh0000gn/T/tmp-10896ZtroXrGxEMe2 --repo https\://charts.jetstack.io
    Error: chart "jetstack/cert-manager" version "v0.10.1" not found in <https://charts.jetstack.io> repository
        at /Users/harrison/code/linio/infrastructure/node_modules/@pulumi/kubernetes/helm/v2/helm.js:112:23
        at OutputImpl.<anonymous> (/Users/harrison/code/linio/infrastructure/node_modules/@pulumi/pulumi/output.js:110:47)
        at Generator.next (<anonymous>)
        at fulfilled (/Users/harrison/code/linio/infrastructure/node_modules/@pulumi/pulumi/output.js:18:58)

cool-egg-852

10/04/2019, 9:23 PM

Just got bit by the API Version changes again. Hopefully support for auto aliasing these will come soon. This time it was in helm though, so out of my control.

millions-judge-24978

10/05/2019, 5:39 PM

I’m seeing an issue with the new k8s auto aliasing. I have a resource of type

kubernetes:autoscaling/v2beta2:HorizontalPodAutoscaler

being created, it’s getting the following aliases that seem to be coming from a different resource altogether:

"aliases": [
                        "urn:pulumi:mirror::mirror::kubernetes:apps/v1:Deployment::traffic-mirror",
                      "urn:pulumi:mirror::mirror::kubernetes:apps/v1beta1:Deployment::traffic-mirror",
                        "urn:pulumi:mirror::mirror::kubernetes:apps/v1beta2:Deployment::traffic-mirror",
                        "urn:pulumi:mirror::mirror::kubernetes:extensions/v1beta1:Deployment::traffic-mirror"
                    ]

resulting in an error like this:

panic: fatal: An assertion has failed: Two resources ('urn:pulumi:mirror::mirror::kubernetes:apps/v1:Deployment::traffic-mirror' and 'urn:pulumi:mirror::mirror::kubernetes:autoscaling/v2beta2:HorizontalPodAutoscaler::traffic-mirror') aliased to the same: 'urn:pulumi:mirror::mirror::kubernetes:apps/v1:Deployment::traffic-mirror'

curved-doctor-83600

10/05/2019, 6:10 PM

Hi, i am new to pulumi and have a question regarding nginx-ingress deployment incl. load balancer (i am trying on digitalocean, if this is of interest). Reading through the docs, using yaml or helm should only be used for experimenting or migrating to pulumi. I am not sure, if I should translate the ingress resources to pulumi code in order to deploy the ingress controller and load balancer? Is this feasible or should I stick to the helm chart/yaml manifest? If so, how would I retrieve the load balancers ip if deployed through a pulumi kubernetes.yaml.ConfigFile?

colossal-room-15708

10/05/2019, 10:39 PM

Does tf2pulumi not support the terraform 0.12 syntax?

error: creating tree module: Error parsing <http://03-main.tf|03-main.tf>: At 7:21: Unknown token: 7:21 IDENT var.ARM_SUBSCRIPTION_ID

damp-room-71337

10/07/2019, 11:07 AM

has anyone got any pointers on namespace handling with the

getResourceProperty

method on Helm charts? I seem to be running into this, although it’s closed as fixed: https://github.com/pulumi/pulumi-kubernetes/issues/419

damp-room-71337

10/07/2019, 11:09 AM

looks like

getResourceProperty

doesn’t like namespace values that are passed as Pulumi a `Output`:

const k8sProvider = new k8s.Provider("k8s-provider", {
  suppressDeprecationWarnings: true,
  namespace: "dev"
});
const port = mysql.getResourceProperty(
  "v1/Service",
  k8sProvider.namespace,
  name,
  "spec"
).ports[0].port;

that raises a

Cannot read property 'spec' of undefined

error, unless I change

k8sProvider.namespace

getResourceProperty

to a string literal - `pulumi.interpolate`/`apply()` doesn’t have an effect either

incalculable-engineer-92975

10/07/2019, 1:15 PM

We've got multiple people working on stacks in the same sandbox. As a result, sometimes resources get messed with/deleted manually while we build out process. But a

pulumi destroy

blows up when resources aren't present to delete. Is there/could there be, a

--force

flag that will ignore resources that aren't present?

Title

incalculable-engineer-92975

10/07/2019, 1:15 PM

We've got multiple people working on stacks in the same sandbox. As a result, sometimes resources get messed with/deleted manually while we build out process. But a

pulumi destroy

blows up when resources aren't present to delete. Is there/could there be, a

--force

flag that will ignore resources that aren't present?

tall-librarian-49374

10/07/2019, 1:22 PM

Have you tried running

pulumi refresh

before

destroy

incalculable-engineer-92975

10/07/2019, 1:31 PM

Yes, it fails all over the place so I end up exporting the stack, deleting ALL resources, and re-importing. Then I delete the underlying EKS cluster and start building it up from scratch.

If I do a

pulumi destroy

, it shouldn't fail on missing resources, I want them gone, so if they're gone already, bonus.

Just give me a notice at the end, 'these resources weren't there to delete', 'these resources couldn't be verified as deleted since the connecting resources were missing'

colossal-beach-47527

10/07/2019, 3:32 PM

@incalculable-engineer-92975 could you please file an issue here? The “warning: went to delete resource, but wasn’t found” seems like a reasonable thing to do for the

pulumi destroy

case. Though ideally a

pulumi refresh

followed by

pulumi destroy

should “just work”. So if you could elaborate on “it fails all over the place”, that would be good too. Since if there are problems with that approach we definitely should look into it and fix them as possible.

white-balloon-205

10/07/2019, 6:25 PM

This has come up several times - and I thought we had an issue but wasn't able to find it either of the last two times it's been raise. I opened https://github.com/pulumi/pulumi/issues/3304 to track. That said - it would be useful to have details on your usecase about what failures you see when you

refresh

then

destroy

(though in general, I can definitely imagine this not working fully if there are fundamental things broken in your stack due to interference outside of Pulumi).

View count: 1