pulumi-community #general

stale-holiday-22431

10/18/2018, 6:25 PM

Are chart problems like these something I usually have to watch out for with Pulumi?

creamy-potato-29402

10/18/2018, 6:26 PM

It’s common for very very complex helm charts to have problems.

creamy-potato-29402

10/18/2018, 6:26 PM

Most of the simple charts will just work out of the box.

stale-holiday-22431

10/18/2018, 6:26 PM

Ah, okay.

creamy-potato-29402

10/18/2018, 6:26 PM

The problems we fix here though have to be fixed anyway for Helm 3.

creamy-potato-29402

10/18/2018, 6:26 PM

Sooo.

creamy-potato-29402

10/18/2018, 6:26 PM

Anyway, our strategy is to upstream fixes when we find bugs in the chart.

stale-holiday-22431

10/18/2018, 6:27 PM

So it’ll all just magically get fixed at some point.

stale-holiday-22431

10/18/2018, 6:27 PM

Cool, I like that strategy.

creamy-potato-29402

10/18/2018, 6:27 PM

Yeah, I’m curious to see what they do with the coreos repository since this is all moving over to charts upstream. I’m hoping the still adopt the fixes in coreos so that new users don’t run into problems. 🙂

creamy-potato-29402

10/18/2018, 6:28 PM

If you find a chart that doesn’t work, come bother me and I’ll fix it and try to upstream the fix.

stale-holiday-22431

10/18/2018, 6:31 PM

What is the

k8sprovider

argument in the code block for?

chilly-pharmacist-32380

10/18/2018, 7:06 PM

I can't answer that question

orange-tailor-85423

10/18/2018, 7:07 PM

following the examples pretty well to get a cluster set up in GKE

orange-tailor-85423

10/18/2018, 7:07 PM

now I'd like to "adopt" one of our helm chart sample apps our devs wrote

orange-tailor-85423

10/18/2018, 7:07 PM

samples all show repo: "stable" etc. Does this just get swapped out for an IP/URL destination for our chart(s) ?

full-dress-10026

10/18/2018, 8:22 PM

I want to deploy an API with API Gateway that has a single resource

/{proxy+}

with an

ANY

method. It will simply send all requests to a lambda function using the

LAMBDA_PROXY

request integration. What is the idiomatic way to go about this? i.e. Should I create all of the individual AWS resources using the aws package or is there some higher level package, perhaps in cloud-aws or cloud, that would allow me to do that with less code? The lambda function is an existing lambda function created and managed by a different, non-pulumi, process.

full-dress-10026

10/18/2018, 8:31 PM

Ah, I see this example is pretty much exactly what I'm after https://github.com/pulumi/examples/blob/master/aws-ts-serverless-raw/index.ts. Is there a reason you wanted to explicitly create the stage? https://github.com/pulumi/examples/blob/0cd79d7c10bd5198f13326d2660d84fbd60fa5ed/aws-ts-serverless-raw/index.ts#L107-L108

busy-umbrella-36067

10/18/2018, 8:33 PM

Copy code

$ cat stack.json | noglob jq -r .deployment.resources[3]
{
  "urn": "urn:pulumi:containers-dev::containers::pulumi:providers:kubernetes::default",
  "custom": true,
  "id": "4bee1431-e9c0-42d0-99f1-e813af0fa1a4",
  "type": "pulumi:providers:kubernetes",
  "inputs": {
    "version": "0.17.2"
  },
  "dependencies": null,
  "initErrors": null
}

$ cat stack.json | noglob jq -r .deployment.manifest.plugins[0]
{
  "name": "kubernetes",
  "path": "/Users/azamat/.pulumi/plugins/resource-kubernetes-v0.17.2/pulumi-resource-kubernetes",
  "type": "resource",
  "version": "0.17.2"
}

$ pulumi plugin ls                                                                             ✘ 255
NAME                       KIND         VERSION                    SIZE               INSTALLED          LAST USED
aws                        resource     0.16.0                     144 MB             2 minutes ago      now
kubernetes                 resource     0.17.2                     48 MB              2 minutes ago      now

TOTAL plugin cache size: 192 MB

$ pulumi refresh

...
...
...

error: could not find plugin for provider 'urn:pulumi:containers-dev::containers::pulumi:providers:kubernetes::default'

full-dress-10026

10/18/2018, 9:38 PM

How do I pass an output to the build args of a

Service

? For example, I have

Copy code

export let apiEndpoint = deployment.invokeUrl.apply(url => url + deployment.stageName);

let nginxService = new cloudAws.Service("nginx", {
    containers: {
        nginx: {
            build: {
                context: "./nginx",
                args: {
                    "API_URL": apiEndpoint,
                    "FRONTEND_URL": frontendUrl
                }
            },
            memory: 256,
            ports: [{port: 443, external: true, protocol: "https"}]
        }
    },
    replicas: 3
});

But it appears build args requires a string, not an

Output

. In the above example,

apiEndpoint

is an output.

full-dress-10026

10/18/2018, 11:01 PM

What should the value of

id

be for the

aws.lambda.Function.get

function?

full-dress-10026

10/18/2018, 11:01 PM

I tried passing my Lambda function's name but got

Copy code

Plan apply failed: refreshing urn:pulumi:services-prod::services::aws:lambda/function:Function::ion-lambda-function: InvalidParameter: 1 validation error(s) found.
    - minimum field size of 1, GetFunctionInput.FunctionName.

full-dress-10026

10/18/2018, 11:03 PM

Or perhaps I shouldn't pass a value for that parameter?

quaint-queen-37896

10/18/2018, 11:07 PM

sup

✋ 3

👋🏼 3

full-dress-10026

10/18/2018, 11:33 PM

Also tried calling it by passing the lambda ARN:

Copy code

let ionLambdaFunction = aws.lambda.Function.get("ion-lambda-function", null,
    {arn: ionLambdaArn}
);

full-dress-10026

10/18/2018, 11:33 PM

which did not work.

full-dress-10026

10/18/2018, 11:40 PM

This works:

Copy code

let ionLambdaFunction = aws.lambda.Function.get("ion-lambda-function", ionLambdaArn,
    {name: ionLambdaName}
);

And this does not:

Copy code

let ionLambdaFunction = aws.lambda.Function.get("ion-lambda-function", null,
    {name: ionLambdaName}
);

Curious why you'd need to pass the lambda ARN and the name.

full-dress-10026

10/18/2018, 11:40 PM

The latter fails with

Copy code

error: Error: Missing required property 'handler'
        at Function (/home/kenny/compute_software/infrastructure/pulumi-src/services/node_modules/@pulumi/lambda/function.ts:171:23)
        at Function.get (/home/kenny/compute_software/infrastructure/pulumi-src/services/node_modules/@pulumi/lambda/function.ts:24:16)
        at Object.<anonymous> (/home/kenny/compute_software/infrastructure/pulumi-src/services/index.ts:49:45)
        at Module._compile (module.js:541:32)
        at Module.m._compile (/home/kenny/compute_software/infrastructure/pulumi-src/services/node_modules/ts-node/src/index.ts:439:23)
        at Module._extensions..js (module.js:550:10)
        at Object.require.extensions.(anonymous function) [as .ts] (/home/kenny/compute_software/infrastructure/pulumi-src/services/node_modules/ts-node/src/index.ts:442:12)
        at Module.load (module.js:458:32)
        at tryModuleLoad (module.js:417:12)
        at Function.Module._load (module.js:409:3)
    error: an unhandled error occurred: Program exited with non-zero exit code: 1

white-balloon-205

10/18/2018, 11:48 PM

For

.get

methods, you always need to provide an

id

, and sometimes need to provide additional property values. What is required here is ultimately based on what the semantics of the underlying Terraform AWS provider

Read

method, which is unfortunately not documented robustly. Most of the AWS resources don't require anything beyond the

id

, but some do. In particular,

aws_lambda_function

in Terraform does use the

name

and optionally the

qualifier

to look up the function instead of using the

id

. That error message doesn't look great though - and I suspect we can report a better error message in this case.

brave-angle-33257

10/19/2018, 12:27 AM

any way to install the old pulumi version temporarily? I upgraded to .16 using the .sh file method, but now I have some hanging resources when the names changed assignment->Assignment.. was hoping to go back to old version and delete them then upgrade and refresh/redeploy