hi, I have a weird issue with ignoreChanges, there's a resource in a ComponentResource class:

const vpc = new aws.ec2.Vpc(
      vpcName,
      {
        cidrBlock: cidrBlock,
        enableDnsHostnames: true,
        enableDnsSupport: true,
        tags: Object.assign(
          {
            Name: vpcName
          },
          args.resourceTags
        )
      },
      { parent: this }
    )

if I change this to

const vpc = new aws.ec2.Vpc(
      vpcName,
      {
        cidrBlock: cidrBlock,
        enableDnsHostnames: true,
        enableDnsSupport: true,
        tags: {
          Name: vpcName
        }
      },
      { parent: this, ignoreChanges: ['tags'] }
    )

and trying to update the stack, I got an error:

aws:ec2:Vpc (dlv-sbx-vpc):
    error: aws:ec2/vpc:Vpc resource 'dlv-sbx-vpc' has a problem: tags (kubernetes): '' expected type 'string', got unconvertible type 'map[string]interface {}'
 
  pulumi:pulumi:Stack (dliver-dlv-sbx):
    error: Running program '/home/ncsibra/dev/prmrgt-infra/pulumi/dliver' failed with an unhandled exception:
    Error: invocation of aws:ec2/getVpc:getVpc returned an error: transport is closing
        at monitor.invoke (/home/ncsibra/dev/prmrgt-infra/pulumi/node_modules/@pulumi/pulumi/runtime/invoke.js:74:33)
        at Object.onReceiveStatus (/home/ncsibra/dev/prmrgt-infra/pulumi/node_modules/grpc/src/client_interceptors.js:1205:9)
        at InterceptingListener._callNext (/home/ncsibra/dev/prmrgt-infra/pulumi/node_modules/grpc/src/client_interceptors.js:568:42)
        at InterceptingListener.onReceiveStatus (/home/ncsibra/dev/prmrgt-infra/pulumi/node_modules/grpc/src/client_interceptors.js:618:8)
        at callback (/home/ncsibra/dev/prmrgt-infra/pulumi/node_modules/grpc/src/client_interceptors.js:845:24)

This resource has a tag, added by EKS, something like

<http://kubernetes.io/cluster/dlv-sbx-cluster-eks-197239a|kubernetes.io/cluster/dlv-sbx-cluster-eks-197239a>: shared

, this tag not exist in pulumi code, because vpc created first, then EKS, so unable to add it from pulumi, but after a refresh, pulumi tries to remove that, but that would cause problems with resources handled by kubernetes. So I tried to add ignoreChanges, but looks like it want to do something with the tags anyway and unable to parse it.

Any reason why pulumi/azure v0.18.1 package is not in npm yet?

Hi 👋

Just wonder where I can find example code for deploying a

golang

function to

GCF

Using

Typescript

preferrably

Is there an example of what goes in here https://pulumi.io/reference/pkg/nodejs/@pulumi/gcp/cloudfunctions/#Function-eventTrigger trying to have the trigger be from google's pubsub?

i’ve now gotten this error 3 times in the last couple days:

Diagnostics:
  pulumi:pulumi:Stack (compute-studio-compute-studio-b-stage-eus2):
    error: transport is closing

is this the host having networking issues? or could this be pulumi related? not seeing any other obvious signs of networking trouble on the host

Hi, could the homebrew formula be updated to 0.17.10?

Is there an elegant way to reference the AWS SDK in a Lambda callback function? Such as

const ingestFunction = new aws.lambda.CallbackFunction("processor", {
  role: lambdaRole,
  callback: async event => {
    const s3 = new aws.sdk.S3({ apiVersion: "2006-03-01" });
    const foo = s3.listBuckets((err, data) => {
      return data;
    });
    console.log(foo);
  }
});

That function, when run just dumps a huge AWS response object

If I change it to this

const ingestFunction = new aws.lambda.CallbackFunction("ingestProcessor", {
  role: lambdaRole,
  callback: async event => {
    const AWS = require("aws-sdk");
    const s3 = new AWS.S3({ apiVersion: "2006-03-01" });
    s3.listBuckets((err, data) => {
      console.log(data.Buckets);
    });
  }
});

It doesn't log anything - I saw this example - https://github.com/pulumi/examples/blob/master/aws-ts-apigateway/index.ts#L24 - and thought you could reference the SDK that way...but maybe I misunderstood that piece

I've tried a few different things, but still can't seem to correctly convert an output from another stack into a string using stack.getOutput + val.apply(v =>

${v}

). Are there any examples on how to do this? I need to get the pure string value so I can use it in a subsequent child_process.exec call.

I'm trying to use the new

pulumi login gs://...

and running into this error whenever I try to do my first

pulumi preview

for the stack:

error: Could not get signed url for stack location: blob (code=Unknown): to use SignedURL, you must call OpenBucket with a valid Options.GoogleAccessID and exactly one of Options.PrivateKey or Options.SignBytes

Anyone gotten the GCP bucket state to work yet?

Hi, I'm working on a proof of concept to use Pulumi in our project (AWS, Event Sourced). I'm getting an error on

pulumi up

, but I don't really know where to go from there:

run pulumi up -s dev
Previewing update (dev):

    pulumi:pulumi:Stack reventless-examples-dev running 
    reventless:CommandTopic CustomerCommandTopic  
    aws:sqs:Queue CustomerCommandTopicQueue  
    aws:sns:Topic CustomerCommandTopicTopic  
    pulumi:pulumi:Stack reventless-examples-dev  (node:5748) ExperimentalWarning: queueMicrotask() is experimental.
    aws:sqs:QueueEventSubscription CustomerAggregateSubscription  
    aws:sns:TopicSubscription CustomerCommandTopicSubscription  
    aws:iam:Role CustomerAggregateSubscription  
    aws:iam:RolePolicyAttachment CustomerAggregateSubscription-7cd09230  
    aws:iam:RolePolicyAttachment CustomerAggregateSubscription-32be53a2  
    pulumi:pulumi:Stack reventless-examples-dev  Error: failed to register new resource CustomerAggregateSubscription [aws:lambda/function:Function]: 2 UNKNOWN: failed to compute asset hash: asset path 'node_modules/reventless' is a directory; try using an archive
    pulumi:pulumi:Stack reventless-examples-dev      at Object.registerResource (/home/chris/Projects/reventless-examples/node_modules/reventless/node_modules/@pulumi/aws/node_modules/@pulumi/pulumi/runtime/resource.js:114:27)

The setup is as following: The first project (called "reventless") contains some `componentResource`s (pure javascript) and is versioned in a private git repository. Reventless is required by the second project, stated in

package.json

(private repo with access token). This seems to work. Troubles arise, when I use

npm link

to develop both projects together. Is there some known issue regarding linked node modules?

i've never had linked modules working either - when pulumi rebuilds the project it ignores the links and uses the filesystem instead

It seems as though if I stop an ec2 instance and start it up again, pulumi retains the original public ip / dns even though it has changed and even after doing

pulumi refresh

is this a bug?

hi, trying to update from aws 0.17.x to 0.18.x without any code change, only depencencies and pulumi tries to recreate my ec2 instances, because something added to ebsBlockDevices by pulumi:

++aws:ec2/instance:Instance: (create-replacement)
                [id=i-05c0e1f10f4f3799c]
                [urn=urn:pulumi:sbx-app::etcd-clusters::Etcd:EtcdCluster$Etcd:EtcdNode$aws:ec2/instance:Instance::etcd3-ec2Node]
              ~ ebsBlockDevices    : [
                  ~ [0]: {
                          + __defaults         : []
                        }
                  ~ [1]: {
                          + __defaults         : []
                        }
                ]
            +-aws:ec2/instance:Instance: (replace)
                [id=i-05c0e1f10f4f3799c]
                [urn=urn:pulumi:sbx-app::etcd-clusters::Etcd:EtcdCluster$Etcd:EtcdNode$aws:ec2/instance:Instance::etcd3-ec2Node]
              ~ ebsBlockDevices    : [
                  ~ [0]: {
                          + __defaults         : []
                        }
                  ~ [1]: {
                          + __defaults         : []
                        }
                ]

Is this normal?

i believe those __defaults changes wont actually result in deployment changes its some strange artifact of updating the stack

i had the same with azure kube cluster, though make sure you're on latest pulumi CLI

trying to use cross-stack references for first time.. I have my VPC stack that has outputs of the subnet IDs:

Outputs:
  + public_subnet_ids: [
  +     [0]: "subnet-00c7ed0ec22cb3ce4"
  +     [1]: "subnet-04327ab075a2ae3b3"
    ]

I’m then trying to pass those into an ALB in another stack:

const vpc = new pulumi.StackReference(`my-vpc/${env}`);
var alb = new aws.applicationloadbalancing.LoadBalancer(alb_name, {
    name: alb_name,
    subnetMappings: vpc.getOutput("public_subnet_ids"),
    enableCrossZoneLoadBalancing: true,
    enableHttp2: true
});

and getting this:

Diagnostics:
  aws:applicationloadbalancing:LoadBalancer (myapp-alb-stage-alb-main):
    error: aws:applicationloadbalancing/loadBalancer:LoadBalancer resource 'myapp-alb-stage-alb-main' has a problem: subnet_mapping.0: expected object, got string
    error: aws:applicationloadbalancing/loadBalancer:LoadBalancer resource 'myapp-alb-stage-alb-main' has a problem: subnet_mapping.1: expected object, got string

seems like an Output/String/Promise issue but not sure what I’m doing wrong here

in the first part, its probably that vpc.getOutput() is type-inferring to "any" so you're not getting a type error for passing a string[] to subnetMappings

ihni what type of object it expects but if you remove the vpc.getOutput() to a variable it might help

Is there a livesite issue going on?

$ pulumi up -v9
Previewing update (my-stack):
error: [500] Internal Server Error

Any updates on state import? This will become important pretty soon for us

Is it possible to update the execution role for a lambda function declared inline via the

awsx

package?

Similarly - is it possible to change the deployment type of an API endpoint (from

REGIONAL

to

PRIVATE

) in

awsx

?

Anybody got an example for retrieving an azure storage account SAS token in python? This here doesn't work.

sas_token = storage.GetAccountSASResult(
    connection_string=storage_account.primary_connection_string,
    https_only=False,
    resource_types=(
        service=False,
        container=False,
        object=True
    ),
    services=(
        blob=True,
        queue=False,
        table=False,
        file=False,
    ),
    start="2019-01-30",
    expiry="2028-03-21",
    permissions=(
        read=True,
        write=False,
        delete=False,
        list=False,
        add=False,
        create=False,
        update=False,
        process=False
    )
)

Anyone having issues setting up an aks cluster in azure? I manually created a service principal for pulumi to run as, it's got all the permissions i can find but i get this when i try to create a new SP with pulumi

SP creating error: Plan apply failed: graphrbac.ApplicationsClient#Create: Failure responding to request: StatusCode=403 -- Original Error: autorest/azure: Service returned an error. Status=403 Code="Unknown" Message="Unknown service error" Details=[{"odata.error":{"code":"Authorization_RequestDenied","date":"2019-05-07T14:22:29","message":{"lang":"en","value":"Insufficient privileges to complete the operation."},"requestId":"6df1d82f-17bb-48da-94d7-8dd03e6a39fa"}}]

Is there a roadmap that the pulumi team can share?

are there are docs/how-tos out there on using lambda@edge with pulumi?

Hi all, I'm using Pulumi via JavaScript with AWS and seeing some strange behavior with the

getAmi

function. (I freely admit this could be due to my lack of experience with JavaScript.) Here's the code in question:

const bionicAmi  = aws.getAmi({
    filter: [
        { name: "name", values: "ubuntu/images/hvm-ssd/ubuntu-bionic-18.04-amd64-server*" },
        { name: "virtualization-type", values: "hvm" },
        { name: "root-device-type", values: "ebs" },
    ],
    mostRecent: true,
    owners: [ "099720109477" ],
});

Using this set of filters with the AWS CLI works perfectly, but here it returns a completely different AMI. What am I missing?