Hi! Any tips on how to supply userData to AWS ECS instances? I've got ecsCluster.createAutoScalingGroup() and when I pass userData in the launchConfiguration, the original ECS-specific userData is lost and my cluster doesn't start

Hi there

I'm getting some troubles with config module http://codebin.herokuapp.com?s=5ea2ba040fce9c0004000001

anyone got this kind of issue ?

How one is supposed to query nested properties with the CLI in a stack output ? I have an output like:

OUTPUT           VALUE
componentProps   {"foo": "something", "bar": "whatever", ...}

And I want to get it like

pulumi stack output componentProps.foo

or

pulumi stack output componentProps:foo

but didn’t find the right way to do it. Obviously I can do a

pulumi stack output componentProps | jq -r '.foo'

but I was wondering if there was a “native” way in

pulumi

CLI to do it (and avoid the

jq

dependency)

I implemented a ComponentResource which creates some other resources including a gcp.container.Cluster and a k8s.Provider from this cluster. Now I'm searching for a way to add this provider to the ComponentResources providers so that all child resources of my ComponentResource will use the k8sProvider. Is that possible or to I have to provide my new ComponentResource as parent and the k8s.Provider as provider to all child resources. This feels a little bit redundant to me. Is there a way to register providers for a ComponentResource?

Is there a way that I can mark a resource to be created only once and not deleted by

pulumi destroy

? I have a cloud endpoint resource that is causing problems when it's deleted.

Can anyone point me in the right direction for information on managing projects and stacks hosted in the cloud? I have an Azure Storage Account hosting a storage container called pulumi that I want to hold multiple projects and allow the use of StackReference between projects. On my first attempt, I ran "pulumi login azblob://pulumi" I could successfully create a new project (with a stack called dev) which set up .pulumi folder in my cloud storage and life was good. I then tried to create a second project, using the same cloud storage, also with a stack called dev I got a warning that a stack with that name already existed. This seems to be because projects aren't stored in subfolders within .pulumi. I could force differences by including project names in stack names but that would get ugly quickly. On my next attempt I ran "pulumi login azblob://pulumi/project1". That initially worked well and I could then create a second project with overlapping stack names by running "pulumi login azblob://pulumi/project2" before creating the project. Each project folder within my cloud storage now has its own .pulumi folder. This is a bit of a pain as I now have to run pulumi login when I switch between projects and I haven't been able to find a way to get StackReference to work so that Project2 can refer to Project1. What is the correct way to do this? Ideally I would like to be able to have a single .pulumi folder with subfolders for each project and their own stacks managed inside them. Also what is the correct way to use a StackReference between cloud hosted projects? Thanks Alan

What's the idea behind ComponentResource.initialize(), ComponentResource.getData() and the generic TData type? A realworld example would also be useful.

Hi!

Hi, I'm looking for cdn custom domain resource in C#? Is that supported?

Hi. I'm getting a new error with Pulumi 2. My existing projects are still able to deploy correctly, but I just installed a fresh new project using Pulumi 2 and the

aws-javascript

template, and every time I try to run

pulumi up

or

pulumi preview

it fails with this error:

error: an unhandled error occurred: Program exited with non-zero exit code: -1
 
    /usr/local/bin/node[9740]: ../src/env.cc:435:virtual node::Environment::~Environment(): Assertion `(base_object_count()) == (0)' failed.
     1: 0x100b8ff9a node::Abort() (.cold.1) [/usr/local/bin/node]
     2: 0x1000832c4 node::FatalError(char const*, char const*) [/usr/local/bin/node]
     3: 0x10008307c node::AppendExceptionLine(node::Environment*, v8::Local<v8::Value>, v8::Local<v8::Message>, node::ErrorHandlingMode) [/usr/local/bin/node]
     4: 0x10003e4bc node::Environment::~Environment() [/usr/local/bin/node]
     5: 0x10003e606 node::Environment::~Environment() [/usr/local/bin/node]
     6: 0x1000b4ac2 node::NodeMainInstance::Run() [/usr/local/bin/node]
     7: 0x100060f13 node::Start(int, char**) [/usr/local/bin/node]
     8: 0x7fff71d33cc9 start [/usr/lib/system/libdyld.dylib]

Even if I remove all of the actual resources so it has nothing to deploy, it will still crash on the preview. Basically, it never gets to the state where it asks me to proceed or not. Any reports of something like this?

Is it not possible to have multiple projects when using the aws backend? When I try to create a second project with a stack named the same as a stack in my first project, I get an error about the stack already existing.

Does anyone have a working codepipeline example in python? I've tried to translate https://www.pulumi.com/docs/reference/pkg/aws/codepipeline/pipeline/ and get a very unhelpful

TypeError: startswith first arg must be bytes or a tuple of bytes, not str

So I guess I'm doing something wrong with the config, but I can't see it (fixed, see thread).

Hello everyone. I learnt about pulumi recently and would like to switch over from terraform. I am a developer who happens to also manage infrastructure. So, I hope that answers why pulumi. I just started working on mmigrating my gcp -> gke infrastructure management from terraform to pulumi using python sdk. The first problem I ran into this evening is that, I use two different gcp service accounts for terrform backend state management and infrastructure management. Looks like I can not do that in pulumi or can I?

Any general tips for sharing high level definitions (e.g. VPC) across multiple repos / projects? E.g. one VPC may have several deployments in it, from different code bases. I think this can simply be done with internal NPM modules. But I am wondering if there is another, maybe a better, way?

is there a way to configure stacks without having to check them into git? for example if I name my aws profile dev but everyone else on my team names it development how do i keep my configs?

Hi. Do cross project references (using StackReference) just allow access to resources in other projects or do they also provide a level of resource dependency awareness i.e. is Pulumi aware of a relationship between projects? For example if I have a simple 2 project setup where 1. Project1 creates a cloud container (e.g. an Azure Resource Group) and outputs its ID 2. Project2 reads in the container ID outputted by Project1 and creates a resource inside it Does Pulumi know about the dependency and refuse to run a "pulumi destroy" against Project1 resources while Project2 is depending on them? In my tests this wasn't the case and I could destroy Project1, breaking Project2 by removing resources it depended on. Is there any way to achieve this dependency for separate projects?

hey, guys did you think to rework the config part, it's really confusing how it works tbh.

What is the url parameter for in

pulumi new

that replaces the template parameter?

Hello, I have some questions on the pulumi 2 migration I didn’t see covered in the migration guide. Is the pulumi v2 cli compatible with running pulumi v1 projects? Can I upgrade the cli and deploy old projects as usual? How would I rewrite the following code to avoid requireOutputSync? I’ve tried casting to

pulumi.Output<string>[]

but this just produces weird type error I’m unable to resolve without casting to

unknown

.

const vpcId = stackRef.requireOutput('vpcId');
const vpcPrivateSubnetIds = <string[]>stackRef.requireOutputSync('vpcPrivateSubnetIds');
const vpcPublicSubnetIds = <string[]>stackRef.requireOutputSync('vpcPublicSubnetIds');
const vpc = awsx.ec2.Vpc.fromExistingIds('vpc', {
  privateSubnetIds: vpcPrivateSubnetIds,
  publicSubnetIds: vpcPublicSubnetIds,
  vpcId,
});

pulumi thinks i'm in non-interactive mode all the time now after upgrading to 2.0

Why would a Kubernetes resource not have a "unique name" if it's the only resource of its kind with that name deployed? I am seeing this error:

kubernetes:<http://networking.k8s.io:NetworkPolicy|networking.k8s.io:NetworkPolicy> prometheus-monitoring  error: Duplicate resource URN 'urn:pulumi:review-payment-manager-fix-prometheus::payment-manager::kubernetes:<http://networking.k8s.io/v1:NetworkPolicy::prometheus-monitoring';|networking.k8s.io/v1:NetworkPolicy::prometheus-monitoring';> try giving it a unique name

There is only one

new k8s.networking.v1.NetworkPolicy('prometheus-monitoring')

call and resource as far as I can tell

oh

I was going to guess that it was a misbehaving helm chart. 🙂

task_definition = ecs.TaskDefinition(
            name,
            family=name,
            cpu="256",
            memory="512",
            network_mode="awsvpc",
            requires_compatibilities=["FARGATE"],
            execution_role_arn=role.arn,
            container_definitions=json.dumps(
                [
                    {
                        "name": "my-app",
                        "image": image,
                        "portMappings": [{"containerPort": port, "protocol": "tcp"}],
                    }
                ]
            ),
            opts=child_opts,
        )

Is there something I need to do to get the two stacks to work together? I don't have this problem with my other networking project stack, which uses python.

👋 New to Pulumi and so far so good. Something that's tripping me up is that I can't seem to get 2 stacks to work together that have a different runtime. I have a typescript stack that's outputting an image:

(venv) ~/***/pulumi/projects/images/config $ pulumi stack output image
***.<http://dkr.ecr.us-east-1.amazonaws.com/jdrake/***-59fb83d:602e111c06b6934013578ad80554a074049c59441d9bcd963cb4a7feccede7a5|dkr.ecr.us-east-1.amazonaws.com/jdrake/***-59fb83d:602e111c06b6934013578ad80554a074049c59441d9bcd963cb4a7feccede7a5>

and then a python stack that's trying to import it then pass it to a Component:

env = pulumi.get_stack()
image_stack = pulumi.StackReference(f"jdrake/***-image-config/{env}")
image = image_stack.get_output("image")
service = Service("***", env, image, port=80)

However, running

pulumi up

gives me:

error: Program failed with an unhandled exception:
    error: Traceback (most recent call last):
      File "/usr/local/bin/pulumi-language-python-exec", line 85, in <module>
        loop.run_until_complete(coro)
      File "/Users/jdrake/.pyenv/versions/3.6.8/lib/python3.6/asyncio/base_events.py", line 484, in run_until_complete
        return future.result()
      File "/Users/jdrake/.pyenv/versions/3.6.8/lib/python3.6/asyncio/tasks.py", line 180, in _step
        result = coro.send(None)
      File "/Users/jdrake/.pyenv/versions/3.6.8/lib/python3.6/site-packages/pulumi/runtime/stack.py", line 81, in run_in_stack
        await run_pulumi_func(lambda: Stack(func))
      File "/Users/jdrake/.pyenv/versions/3.6.8/lib/python3.6/site-packages/pulumi/runtime/stack.py", line 34, in run_pulumi_func
        func()
      File "/Users/jdrake/.pyenv/versions/3.6.8/lib/python3.6/site-packages/pulumi/runtime/stack.py", line 81, in <lambda>
        await run_pulumi_func(lambda: Stack(func))
      File "/Users/jdrake/.pyenv/versions/3.6.8/lib/python3.6/site-packages/pulumi/runtime/stack.py", line 104, in __init__
        func()
      File "/usr/local/bin/pulumi-language-python-exec", line 84, in <lambda>
        coro = pulumi.runtime.run_in_stack(lambda: runpy.run_path(args.PROGRAM, run_name='__main__'))
      File "/Users/jdrake/.pyenv/versions/3.6.8/lib/python3.6/runpy.py", line 280, in run_path
        run_name, mod_spec, pkg_name).copy()
      File "/Users/jdrake/.pyenv/versions/3.6.8/lib/python3.6/runpy.py", line 85, in _run_code
        exec(code, run_globals)
      File "./__main__.py", line 8, in <module>
        service = Service("***", env, image, port=80)
      File "/Users/jdrake/talos/pulumi/components/service/src/ns_pulumi_component_service/component.py", line 86, in __init__
        "portMappings": [{"containerPort": port, "protocol": "tcp"}],
      File "/Users/jdrake/.pyenv/versions/3.6.8/lib/python3.6/json/__init__.py", line 231, in dumps
        return _default_encoder.encode(obj)
      File "/Users/jdrake/.pyenv/versions/3.6.8/lib/python3.6/json/encoder.py", line 199, in encode
        chunks = self.iterencode(o, _one_shot=True)
      File "/Users/jdrake/.pyenv/versions/3.6.8/lib/python3.6/json/encoder.py", line 257, in iterencode
        return _iterencode(o, 0)
      File "/Users/jdrake/.pyenv/versions/3.6.8/lib/python3.6/json/encoder.py", line 180, in default
        o.__class__.__name__)
    TypeError: Object of type 'Output' is not JSON serializable
    error: an unhandled error occurred: Program exited with non-zero exit code: 1

This is failing when I try to use that image value in a task definition:

That

Object of type 'Output' is not JSON serializable

suggests to me you need an 'apply' somewhere https://www.pulumi.com/docs/intro/concepts/programming-model/#outputs

Oh, wow. Nice work!

I need to populate integration request/response templates in api gateway and I keep them in separate folder as vtl files is there a way to read content of file asset as string? or should I opt for fs operations to read them?