pulumi-community #general

jolly-camera-35709

10/20/2020, 5:48 AM

I thought virtual env is to isolate the runtime env from the local env, in this case, it's not

bitter-application-91815

10/20/2020, 1:36 PM

Hi there,

bitter-application-91815

10/20/2020, 1:36 PM

When i'm redploying fargate cluster with it's tasks, (with simple env var updates), I now get this error

bitter-application-91815

10/20/2020, 1:36 PM

Copy code

aws:ec2:RouteTable (axiom-cloud-private-subnet-3-routetable-staging-g):
    error: 1 error occurred:
    	* updating urn:pulumi:axiom-cloud::staging::aws:ec2/routeTable:RouteTable::axiom-cloud-private-subnet-3-routetable-staging-g: Error creating route: RouteAlreadyExists: The route identified by 0.0.0.0/0 already exists.
    	status code: 400, request id: c7800e9a-2e62-4efb-ad81-34b27f11c628

bitter-application-91815

10/20/2020, 1:37 PM

It should not have to touch subnet routing as it was setup already for those tasks.

bitter-application-91815

10/20/2020, 1:37 PM

i'm just redeploying ...

bitter-application-91815

10/20/2020, 1:40 PM

Copy code

View Live: <https://app.pulumi.com/axiom/staging/axiom-cloud/previews/a92a225e-8e64-40a2-b968-08f187e38f2f>

     Type                                         Name                                               Plan        Info
     pulumi:pulumi:Stack                          staging-axiom-cloud                                            
 ~   ├─ aws:ec2:RouteTable                        axiom-cloud-private-subnet-3-routetable-staging-g  update      [diff: ~routes]
 ~   ├─ aws:ec2:RouteTable                        axiom-cloud-private-subnet-1-routetable-staging-g  update      [diff: ~routes]
 ~   ├─ aws:ec2:RouteTable                        axiom-cloud-private-subnet-2-routetable-staging-g  update      [diff: ~routes]
 +-  ├─ aws:ecs:TaskDefinition                    monitors-staging-g-ecs-task                        replace     [diff: ~containerDefinitions]
 ~   ├─ aws:ecs:Service                           monitors-staging-g-ecs-service                     update      [diff: ~taskDefinition]
     └─ kubernetes:<http://helm.sh/v3:Chart|helm.sh/v3:Chart>               axiomdb-charts-staging-g                                       
        └─ kubernetes:extensions/v1beta1:Ingress  axiomdb/axiomdb-ingress                                        1 warning
 
Diagnostics:
  kubernetes:extensions/v1beta1:Ingress (axiomdb/axiomdb-ingress):
    warning: extensions/v1beta1/Ingress is deprecated by <http://networking.k8s.io/v1beta1/Ingress|networking.k8s.io/v1beta1/Ingress> and not supported by Kubernetes v1.20+ clusters.
 

Do you want to perform this update? yes
Updating (axiom/axiom-cloud)

View Live: <https://app.pulumi.com/axiom/staging/axiom-cloud/updates/67>

     Type                   Name                                               Status                  Info
     pulumi:pulumi:Stack    staging-axiom-cloud                                **failed**              1 error
 ~   ├─ aws:ec2:RouteTable  axiom-cloud-private-subnet-2-routetable-staging-g  updated                 [diff: ~routes]
 ~   ├─ aws:ec2:RouteTable  axiom-cloud-private-subnet-3-routetable-staging-g  **updating failed**     [diff: ~routes]; 1 error
 ~   └─ aws:ec2:RouteTable  axiom-cloud-private-subnet-1-routetable-staging-g  updated                 [diff: ~routes]

bitter-application-91815

10/20/2020, 1:40 PM

of course it's not preventing the ecs-task be updated which is basically all i'm trying to do

brash-student-27683

10/20/2020, 2:49 PM

Hi everyone, I am trying to set up vpc peering with pulumi (crosswalk)

Copy code

vpc1 = new awsx.ec2.Vpc(..)
vpc2 = new awsx.ec2.Vpc(..)

new aws.ec2.VpcPeeringConnection(..)

now that the peering is set up, I need to add the route table entries to both the vpc’s route tables. But with crosswalk, these route tables are internally created for me. How do I access these route tables to add a new route to them? (to point the vpcs to each other)

careful-nail-88325

10/20/2020, 4:29 PM

I'm working on setting up a CD pipeline for Pulumi and I'd like my AWS provider to assume different roles depending on whether I'm running a preview on a branch or an update on master. I'm using

pulumi.runtime.isDryRun()

to determine whether a preview or update is being executed and assuming the corresponding role. Unfortunately, the change in the provider also affects the state of my stack. Is there a recommended way to do this? Currently, I think I'm going to have to abandon the assumeRole functionality of the AWS provider and assume the role with the AWS CLI in advance before executing preview or update.

tall-needle-56640

10/20/2020, 4:32 PM

Getting 404 error, when resource exists. In the below output (clipped for brevity), you can see that the app plan must be the same for both functions, but then it can't find the app plan to update it:

> pulumi up

Previewing update (dev)

Type Name Plan

pulumipulumiStack PulumiPort-dev

+ ├─ azureappserviceFunctionApp app create

- └─ azureappserviceFunctionApp test20201910func delete

Resources:

+ 1 to create

- 1 to delete

2 changes. 5 unchanged

Do you want to perform this update? yes

Updating (dev)

Type Name Status Info

pulumipulumiStack PulumiPort-dev failed 1 error

+ └─ azureappserviceFunctionApp app creating failed 1 error

Diagnostics:

azureappserviceFunctionApp (app):

error: web.AppsClient#CreateOrUpdate: Failure sending request: StatusCode=404 -- Original Error: Code="NotFound" Message="Cannot find ServerFarm with name test20201910-aspa9aae18b." Details=[{"Message":"Cannot find ServerFarm with name test20201910-aspa9aae18b."},{"Code":"NotFound"},{"ErrorEntity":{"Code":"NotFound","ExtendedCode":"51004","Message":"Cannot find ServerFarm with name test20201910-aspa9aae18b.","MessageTemplate":"Cannot find {0} with name {1}.","Parameters":["ServerFarm","test20201910-aspa9aae18b"]}}]

worried-queen-62794

10/20/2020, 9:50 PM

How can I use an output property in a function like

aws.ec2.get_network_interface

which doesn't define it's properties as inputs but rather the raw types, i.e. str?

worried-queen-62794

10/20/2020, 10:39 PM

Is there a way to force an output to be stable? It's not my provider so I can't edit it there.

worried-queen-62794

10/20/2020, 10:43 PM

Alternatively, if the preview says that there is going to be a replacement because of an output property but that output property doesn't actually change will the resource be replaced or not?

little-cartoon-10569

10/20/2020, 11:00 PM

Does anyone know if there is a bash prompt for Pulumi? I'd love to add the name of the current stack into my bash-git-prompt code...

brief-spoon-92183

10/21/2020, 11:37 AM

Hi everyone, is there a way to reference a resource ID within

itself

? (here's an example) I'm creating an Elastic Search Domain, and the

accessPolicies

requires a

Resource

ARN, which I don't know how to construct to match the domain ID (see screenshot):

billions-forest-38488

10/21/2020, 12:52 PM

Hi - does anyone know how I can add additional tasks to a

pulumi destroy

command, for example: only when destroying a stack, I'd also like to run a bit of code to cleanup some third-party resources... in this case, cleanup a Logentries logset/log - thanks! (I'm using Python if that helps..)

high-morning-17948

10/21/2020, 9:26 PM

Hi, I'm having issues creating a jobDefinition for a batch process. I keep getting the following error:

Copy code

aws:batch:JobDefinition (booksloth-import-dev-importLauncher-job-definition):
    error: 1 error occurred:
        * : Error executing request, Exception : The validated object is null, RequestId: 3b39ec6f-fcaf-4ea9-bb13-52991d45c6c1
        status code: 400, request id: 3b39ec6f-fcaf-4ea9-bb13-52991d45c6c1 "booksloth-import-dev-importLauncher-job-definition"

Any suggestion of what I'm missing for the creation. Here's the code snippet for creating the resource

Copy code

/**
     * Define Job definition Container properties
     * See {@link <https://docs.aws.amazon.com/batch/latest/APIReference/API_ContainerProperties.html> |  AWSContainerProperties}
     * for reference
     */
    const containerProperties = jobRole.arn.apply(arn => {
      return JSON.stringify({
        image: imgName,
        command: ['node', 'index.js'],
        memory: 2000, // in MiB
        jobRoleArn: arn,
        environment: [
          { name: 'NODE_ENV', value: config.stage },
          { name: 'APP_VARS', value: appVars },
          { name: 'BOOKSLOTH_API_VARS', value: bookslothApiVars },
        ],
      })
    });

    console.log({ containerProperties })
    /**
     * Create Batch Job Defintion
     */
    const jobDefinition = new aws.batch.JobDefinition(`${config.prefix}-${serviceName}-job-definition`, {
      name: `${config.prefix}-${serviceName}-job-definition`,
      type: 'container',
      containerProperties,
      retryStrategy: { attempts: 3 }
    }, {
      dependsOn: [
        jobRoleECSPolicyAttachment,
        jobRoleBookQueueAccess,
        jobRoleUserbooklistQueueAccess
      ]
    });

🙏 1

famous-area-1376

10/22/2020, 6:25 AM

Hi, I'm trying to evaluate if pulumi is a good tool for our company use. I have a simple usecase where I want to 1) Create ECR in AWS, 2) Push a container image to said ECR, 3) Create AWS Batch Job Definition (+Job Queue) based on the image pushed to ECR. I tried doing this with the Python version but I can't figure out how to do it. There's this for node.js: https://www.pulumi.com/blog/building-and-publishing-docker-images-to-a-private-amazon-ecr-repository/ but I can't find the matching tools for Python. So the issue I have is specifically on the part how to push the image to ECR and get reference to that. I think I know how to create the AWS resources, but the image needs to exist in the ECR first before I can create the Job Definition since it requires a reference to it. Only way is node.js or am I missing something in Python tooling?

jolly-camera-35709

10/22/2020, 6:52 AM

anybody knows how to join string and resource output in python with pulumi?

jolly-camera-35709

10/22/2020, 6:54 AM

or in general, how to join string and resource output as input to other resource?

jolly-camera-35709

10/22/2020, 6:55 AM

for example, when I create container group, I wanted to set image like this

jolly-camera-35709

10/22/2020, 6:57 AM

image='{}/{}-build:{}'.format(container_registry.login_server, project_name, tag)

jolly-camera-35709

10/22/2020, 6:58 AM

in python, it's simply a format to join strings, but here it uses the output from container_registry, so the image became "<pulumi.output.Output object at 0x7fb1d850a550>/myproject:202010191642\n"

breezy-butcher-78604

10/22/2020, 3:15 PM

is it possible to be logged to pulumi (via the CLI) with two different accounts? eg i'd like to be able to switch between my work account and personal account but running

pulumi login

again just logs me in with the existing account (my work one)

brief-spoon-92183

10/22/2020, 5:02 PM

Has anyone tried deploying multiple Elastic Search domains? How did you solve the problem of existing

AWSServiceRoleForAmazonElasticsearchService

? The error looks like this:

Copy code

Error creating service-linked role with name <http://es.amazonaws.com|es.amazonaws.com>: InvalidInput: Service role name AWSServiceRoleForAmazonElasticsearchService has been taken in this account, please try a different suffix.

I followed the

VPC based ES

example in the docs. Any ideas how to either reuse the existing role, or create a new one for each ES domain? Thanks

many-spring-73557

10/22/2020, 6:43 PM

I am creating a

gcp.cloudrun.DomainMapping

. Every time I

pulumi up

, it says there’s a change to metadata annotations

serving.knative.dev/creator

and

serving.knative.dev/lastModifier

which match my username. This results in GCP destroying and re-creating the DomainMapping each time, which I don’t want. How can I keep this from happening? One thing I’ve found is I can

ignoreChanges: ["metadata.annotations"]

which is pretty good - but I do have some annotations that should actually get picked up by pulumi. I haven’t found a way to target those specific annotations that I want to ignore

breezy-butcher-78604

10/23/2020, 4:51 AM

i have the following scenario and am looking for advice on the best way to proceed. • I have a stack containing some EC2 instances and associated security groups, roles and an ALB etc • I applied a stack update that was going to replace the vast majority of the resources mentioned above (moving them into a different VPC) • The update failed after a few of the replacement resources were created in the new VPC (for a valid reason) • The following update is now trying to "complete deletion from previous update", in particular a security group, but it is failing because the group still has the old ALB in it (which would have been deleted in the first update had it succeeded). • I now cannot apply any updates to this stack since changes will only occur after completing the previous deletes is there a way I can tell pulumi to not try completing the pending deletes until after the rest of the stack has updated? is there another approach i can use to fix this stack? I'd rather not have to destroy and recreate it if I can avoid it

jolly-camera-35709

10/23/2020, 6:14 AM

hey guys, how do you make this

jolly-camera-35709

10/23/2020, 6:14 AM

environment_variables Mapping[str, str]