pulumi-community #python

mysterious-hamburger-19066

05/25/2022, 6:31 PM

Hello. How do I copy an entire directory from local machine to the VM created by pulumi? I see that there is a command plugin, but it only seems to allow you to copy a file.

quick-telephone-15244

05/26/2022, 6:38 PM

complete and utter shot in the dark--has anyone ever seen something similar to this error when creating an EKS cluster with ManagedNodeGroups

debug_error_string = "{"created":"@1653589486.232131000","description":"Error received from peer ipv4:127.0.0.1:64115","file":"src/core/lib/surface/call.cc","file_line":904,"grpc_message":"Cannot read properties of undefined (reading 'map')","grpc_status":2}"

quick-telephone-15244

05/26/2022, 6:40 PM

literally at

-v=9

for

up

and it's still not entirely clear how i'm getting to that/what the underlying issue is.

purple-architect-30534

05/27/2022, 7:40 PM

I'm getting this using pulumi in Github Actions and I'm not sure how to fix this

orange-crowd-9665

06/01/2022, 9:34 AM

Hello! On GCP, I'm creating PubSub Schemas (w/ gRPC), and then after PubSub topics with those schemas (w/ ResourceOptions depends_on parameter). A problem arises when I update my .proto files. I don't know the inners of Pulumi, but everytime I update a schema, the PubSub topics that rely on it will result in a "Deleted Schema" state. What are the best practices to update PubSub schemas?

crooked-sunset-90921

06/02/2022, 3:42 AM

having a brain block... if I have a function that returns a dict, how do I get just the one specific key from the function... it's obviously not myfunc().get("bla"), but something on those lines?

white-terabyte-21934

06/02/2022, 7:36 AM

Hello , Looking for an advice , I am trying pulumi with python and would like to know the better usage of Output . • Creating a resources let say r = xx.CreateResources(…) ==> r (A pulumi output object) • Would like to extract the raw value of a property ,let say r.http_url (which is again a type of

pulumi.output.Output object

. • Would like to print these values via a print statement (like terraform output format) ,what is the best way to fetch the value via Ouput() method ?

breezy-book-15761

06/02/2022, 9:43 AM

Hello, I'm trying to set up a gcp project from scratch, including enabling some APIs and I'm trying to set up some sort of dependency control to get it to work:

Copy code

"""A Google Cloud Python Pulumi program"""

import pulumi
import pulumi_gcp as gcp

cloud_resource_manager_api = gcp.projects.Service('crm_api', service="<http://cloudresourcemanager.googleapis.com|cloudresourcemanager.googleapis.com>")
project = gcp.organizations.get_project_output(opts=pulumi.InvokeOptions(parent=cloud_resource_manager_api))
cloud_run_api = gcp.projects.Service('cloud_run_api', disable_dependent_services=True, disable_on_destroy=True, project=project.id.apply(lambda project_id: project_id), service="<http://run.googleapis.com|run.googleapis.com>")

salmon-art-25154

06/02/2022, 6:52 PM

hey all! i'm getting an error when running

pulumi update

and i've been trying to solve it since yesterday. the error is:

AttributeError: 'Service' object has no attribute 'status'

full stack trace (https://pastebin.com/jS2FZpPK) initially, i was experiencing the issue described here https://github.com/pulumi/pulumi/issues/9694, which i was able to fix by pinning protobuf v3.20.1 after that though, i began experiencing the error i pasted above. i've tried downgrading pulumi as well as upgrading python (this app uses 3.9.6), but i get the same error no matter what. anyone else experience the same issue or have any tips to help me further debug this?

salmon-art-25154

06/02/2022, 7:34 PM

my code looks like `

Copy code

pulumi.export('api_url', cloud_run_api_output.status.url)

but when i look at the docs for gcp.cloudrun.Service, i only see output.statuses, so maybe that changed between versions. however when i try to access

cloud_run_api_output.statuses[0]

, i get a index out of range error.

salmon-art-25154

06/02/2022, 8:22 PM

well i did it again and statuses was not empty that time. not sure why it was empty the first, but ill take it!

microscopic-holiday-73461

06/04/2022, 5:20 AM

Hey, I just got around to trying out Pulumi for the first time, and I seem to have some issues with the Hetzner integration. Specifically, when I try to create a server bound to a placement group and a network, I get this:

Copy code

Diagnostics:
  hcloud:index:Server (master-1):
    error: hcloud:index/server:Server resource 'master-1' has a problem: Attribute must be a whole number, got 47074. Examine values at 'Server.PlacementGroupId'.
    error: hcloud:index/server:Server resource 'master-1' has a problem: Attribute must be a whole number, got 1710474. Examine values at 'Server.Networks'.

Seems that the Server object expects IDs to come in as integers, but the objects provide IDs as strings. Here's the relevant parts form the code:

Copy code

import pulumi_hcloud as hcloud

network = hcloud.Network("cluster-net",
  ip_range="10.0.1.0/24",
)
master_group = hcloud.PlacementGroup("masters", type="spread")

for i in range(3):
    node = hcloud.Server(
        f"master-{i+1}",
        backups=False,
        location="hel1",
        server_type="CX21",
        image="ubuntu-20.04",
        networks=[{
            "network_id": network.id,
        }],
        placement_group_id=master_group.id
    )
    hcloud.Rdns(
        f"master-{i+1}",
        server_id=node.id,
        ip_address=node.ipv4_address,
        dns_ptr=f"master-{i+1}.<http://kube.golyalpha.tk|kube.golyalpha.tk>"
    )

The network and placement group gets created just fine beforehand

mysterious-hamburger-19066

06/04/2022, 9:38 PM

Has anyone here set up a ray cluster with pulumi? https://www.ray.io/

crooked-sunset-90921

06/05/2022, 2:05 AM

Does anyone co-mingle pulumi code within other python project folders? Especially working within monorepos. Looking for some best practices how to structure folders

victorious-exabyte-70545

06/06/2022, 7:45 PM

Hi all, I am trying to ignore a helm chart version change.

victorious-exabyte-70545

06/06/2022, 7:46 PM

Copy code

rabbitmq_chart = Chart(
    'rabbitmq-chart',
    ChartOpts(
        resource_prefix=stack_name,
        chart='rabbitmq',
        version="8.16.2",

victorious-exabyte-70545

06/06/2022, 7:46 PM

8.16.2 is not available at the moment so I simply want to ignore version in chartopts.

victorious-exabyte-70545

06/06/2022, 7:48 PM

I tried this transformation:

Copy code

def transformation(args: ResourceTransformationArgs):
    if args.type_ == "kubernetes:helm:template":
        return ResourceTransformationResult(
            props=args.props,
            opts=ResourceOptions.merge(args.opts, ResourceOptions(
                ignore_changes=["ChartOpts"],
            )))

victorious-exabyte-70545

06/06/2022, 7:49 PM

But it still fails.

victorious-exabyte-70545

06/06/2022, 7:49 PM

Any ideas on how I can ignore version there

damp-fall-55922

06/07/2022, 8:46 PM

I have a weird issue and I don’t know if this goes in #aws, here, or elsewhere but here goes. I built a stack in python that mimics what we currently do with Cloudformation to deploy a Fargate app, reusing outputs from a base networking Cloudformation stack that provides VPC, subnets, load balancers, listeners, DNS zone etc… I built it using a combination of

aws

and

aws-native

providers, preferring the native one where available and working. it works fine, outperforms cloudformation as expected I then moved that exact same code into a new file and wrapped it in

class MyCustomFargateApp(pulumi.ComponentResource):

I passed `opts=pulumi.ResourceOptions(parent=self)`everywhere, and that’s all the changes I did when I tried to deploy it, I got this error:

Copy code

Diagnostics:
  pulumi:providers:aws-native (default_5_7_0):
    error: no resource plugin 'pulumi-resource-aws-native' found in the workspace at version v5.7.0 or on your $PATH, install the plugin using `pulumi plugin install resource aws-native v5.7.0`

which makes no sense because

aws-native

is in v0.18.0 and it’s

aws

that’s v5.7.0 to debug: • I nuked venv and

__pycache__

• I changed the name of the stack in Pulumi.yaml • I removed both providers using

pulumi plugin rm resource aws

etc the results are the same I did the same steps with the working stack without a ComponentResource, it deploys fine from scratch I repeated the same steps with my broken stack but I commented out *every resource*… it deploys fine, just shows an empty MyCustomFargateApp resource. swell. I then proceeded to remove comments from resources one by one. working fine: • pulumi_aws.route53.Record • pulumi_aws.iam.Role • pulumi_aws_native.logs.LogGroup • pulumi_aws_native.ecs.TaskDefinition • pulumi_aws.lb.TargetGroup it blows up when I uncomment the pulumi_aws_native.elasticloadbalancingv2.ListenerRule it does reference the TargetGroup which comes from aws-classic… but exact same stuff works if not wrapped in a ComponentResource please advise, I’m losing my mind

kind-jelly-61624

06/08/2022, 10:49 PM

Is there a way to unwrap and import the Inputs and Properties of a resource from an external schema.json into your own? I’m building a ComponentResource that’s a wrapper over a pulumi-aws resource. I’d like the Component to accept the same inputs, and return the same outputs as the pulumi-aws resource - but I also don’t want to add each input and output to the schema.json manually. Is there a way to import a resource directly into my schema.json?

damp-fall-55922

06/10/2022, 5:25 PM

I’m building a provider in python… was wondering if there’s a shortcut to creating the ComponentNameArgs class? it seems like it could be generated straight from the schema.json, no? in fact, in the python sdk it is generated. do I just copy it back from the SDK to the source? 😄

damp-fall-55922

06/10/2022, 6:31 PM

https://github.com/pulumi/pulumi-component-provider-py-boilerplate/pull/17

green-motorcycle-58080

06/11/2022, 12:59 PM

If I have a dynamic provisioner that is a long running bash script, does

pulumi up

wait until it's finished? If not, what's the best way to track progress of this script?

creamy-whale-55909

06/12/2022, 5:55 PM

I've been experiencing some inconsistency with grabbing the async output values using

pulumi.Output.all()

apply()

. My observed behavior is sometimes, instead of getting access to the output value in it's expected form (ie

str

), I am given the output object address memory (ie

<pulumi.output.Output object at 0x7f02baf74b80>

). I've seen a few folks in the channel struggling with the same problem. I think I understand why this is happening. Sharing my understanding below so folks can

a)

confirm (or correct) my understanding and

b)

benefit from the explanation (assuming it's correct 😅 ). Here's the context... 🧵

worried-gold-55244

06/13/2022, 10:20 PM

Hi, I am not able to update a global variable from within the function(I defined it as global inside), did anybody experience the same ?

quick-wolf-8403

06/14/2022, 2:32 AM

Possibly dumb question: I'm adding Pulumi to a co-worker's Flask-based API. Would it be appropriate to put Automation API stuff in the module's dunder init file?

__init__.py

lemon-processor-10785

06/14/2022, 8:04 PM

When creating a vpc resource in AWS Crosswalk (awsx), is it possible to iterate over the output values in vpc.private_subnet_ids to create a custom route in each of the route tables? I've been trying a few different apply/lambda/Output iterations, and I can't get anything to actually create the new routes.

Copy code

def set_route_tables(subnetIds):
    for subnetId in subnetIds:
        # Get the current routeTable for the subnet. Pulumi creates a new table for each subnet via awsx in vpc creation
        routeTable=aws.ec2.get_route_table(subnet_id=subnetId)
        
        # Create the route to MongoDB Atlas
        mongoRoute=aws.ec2.Route(env + "-mongoroute-" + subnetId, 
            route_table_id=routeTable.id,
            destination_cidr_block=mongoContainer.results[0].atlasCidrBlock,
            vpc_peering_connection_id=mongoPeering.connectionId,
        )

vpc.private_subnet_ids.apply(lambda subnetIds: set_route_tables(subnetIds))

👍 1

creamy-whale-55909

06/15/2022, 12:36 AM

I have created a component resource that makes a call to

get_caller_identity()

in order to grab the current account id of the locally assumed aws role so it can be used to define an arn in a policy doc: (code snippet has been truncated to save space)

Copy code

import pulumi_aws as aws

# gets account id
current = aws.get_caller_identity()
account_id = current.account_id

# passes account id when defining policy
self.lambda_policy = aws.iam.RolePolicy(
            "lambdaPolicy",
            role=self.lambda_role.name,
            policy=self.my_lambda.name.apply(
                lambda physical_name: json.dumps(
                    {
                        "Version": "2012-10-17",
                        "Statement": [
                            {
                                "Effect": "Allow",
                                "Action": "logs:CreateLogGroup",
                                "Resource": f"arn:aws:logs:us-east-1:{account_id}:*"

...

When I instantiate this component resource in my project

__main__.py

, it works as expected. However, when I run

pytest

, I get the following error:

Copy code

________________________________ ERROR collecting tests/test_ip_scan.py _________________________________
tests/test_ip_scan.py:18: in <module>
    test_args = ip_scan.MonitorAvailableIpsArgs(
<string>:9: in __init__
    ???
components/ip_scan.py:28: in __post_init__
    account_id = aws.get_caller_identity().account_id
.venv/lib/python3.10/site-packages/pulumi_aws/get_caller_identity.py:106: in get_caller_identity
    account_id=__ret__.account_id,
E   AttributeError: 'NoneType' object has no attribute 'account_id'

Any and all feedback is greatly appreciated!