pulumi-community #general

general

early-musician-41645

07/24/2019, 7:12 PM

x-posting from a thread in case anyone has experience with this: https://pulumi-community.slack.com/archives/C84L4E3N1/p1563995505114100?thread_ts=1563577662.042800&cid=C84L4E3N1

chilly-photographer-60932

07/24/2019, 7:47 PM

How do I compare

Output<string>

with

string

? Here is the example

const clusterName = xyz
                .getOutput('cluster')
                .apply(cluster => cluster.ClusterName) as Output<string>;

The code has to compare the

clusterName

. I have tried with something like this and it failed.

switch (clusterName) {
            case Output.create('alpha'):{

early-musician-41645

07/24/2019, 8:31 PM

Another secrets question: how can I test to see if a secret exists with a specific name? Should I use

aws.secretsmanager.getSecret({name: "foo"});

with a try-catch?

important-leather-28796

07/24/2019, 8:39 PM

I’ve got an infrastructure diff between osx and linux on our infrastructure stack running the same code, same cli version. The linux execution (ci node) wants to replace our

provider

- making this a bad situation. We

pulumi refresh

on osx yesterday, my understanding is that would affect osx identical to our linux ci node. Who can help with this production issue?

plain-businessperson-30883

07/24/2019, 9:49 PM

Hi. Has anyone ever tried to remove an resource from a helm chart? I’m trying to install the

stable/grafana

chart but it is installing the tests too. I’d like to exclude them, but I can’t find a way to do it. Any suggestions?

swift-painter-31084

07/24/2019, 11:55 PM

I'm having a compile-time error that just started happening after I destroyed my stack and tried to recreate:

Invoke of 'aws:sqs/getQueue:getQueue' failed: undefined (undefined)

const queue = new aws.sqs.Queue( `${ config.namespace }_${ resourceName }` );

// get the queue's url
const queueUrl = pulumi.output( aws.sqs.getQueue( {
    name: queue.name,
} ) ).apply( v => v.url );

// get the queue's arn
const queueArn = pulumi.output( aws.sqs.getQueue( {
    name: queue.name,
} ) ).apply( v => v.arn );

module.exports = () => queue;
module.exports.arn = queueArn;
module.exports.url = queueUrl;

full-dress-10026

07/25/2019, 12:00 AM

I am trying to create a Secret using Pulumi k8s:

new k8s.core.v1.Secret("foo-test", {
    metadata: {
        name: "foo-test"
    },
    data: {
        key: "asd123"
    },
    type: "Opaque"
}, {provider: k8sProvider});

and have been getting

error: Plan apply failed: Secret in version "v1" cannot be handled as a Secret: v1.Secret.ObjectMeta: v1.ObjectMeta.TypeMeta: Kind: Data: decode base64: illegal base64 data at input byte 4, error found in #10 byte of ...|":"asd123"},"kind":"|..., bigger context ...|{"apiVersion":"v1","data":{"key":"asd123"},"kind":"Secret","metadata":{"labels":{"app.kuber|...

I can create the secret from the CLI just fine.

full-dress-10026

07/25/2019, 12:44 AM

If I want to modify a nested yaml string (e.g.

|-

) in a transformation (https://github.com/pulumi/pulumi-kubernetes/blob/4e61e0d307317a57c249ba599d206e5703f89be1/tests/integration/yaml-url/step1/index.ts#L22-L37), is there a way to easily do that?

full-dress-10026

07/25/2019, 12:53 AM

Are there yaml parse and unparse utils in pulumi?

damp-notebook-63293

07/25/2019, 1:56 AM

@full-dress-10026 function in

transformation

receives given yaml string already parsed to JS object. given your yaml string, it should look like this:

{
  "apiVersion": "v1",
  "kind": "ConfigMap",
  "metadata": {
    "name": "honeycomb-agent-config",
    "namespace": "default"
  },
  "data": {
    "config.yaml": "# By default, submit logs from interesting system pods such as the\n# kube DNS\nwatchers:\n  - dataset: kubernetes-logs\n    labelSelector: \"k8s-app=aws-node\"\n    namespace: kube-system\n    parser: glog\n  - dataset: kubernetes-logs\n    labelSelector: \"k8s-app=kube-proxy\"\n    namespace: kube-system\n    parser: glog\n  - dataset: kubernetes-logs\n    labelSelector: \"k8s-app=kube-dns\"\n    namespace: kube-system\n    parser: glog\nverbosity: info\n# splitLogging directs trace, debug, info, and warn log levels to stdout\n# rather than stderr.\nsplitLogging: false"
  }
}

therefore, you need to

JSON.parse(manifest.data["config.yaml"])

, modify the object as you need, and

JSON.stringify()

it back to

manifest.data["config.yaml"]

billions-lock-80282

07/25/2019, 9:45 AM

Hi, I'm trying to deploy nginx ingress controller helm chart in EKS. I'm sticking with the standard nginx controller for now as am sharing the same component across all 3 major clouds. In EKS I seem to have an issue with the cluster IP. Because no cluster IP and LB IP is provided, it generates its own clusterIPs. Everytime I run pulumi up, it tries to remove the assigned IPs as they're set to "" in the manifests. It then complains that the ClusterIP is immutable. Any work around?

gray-lawyer-89054

07/25/2019, 10:30 AM

Hi, Is there any way to merge configurations, so that I can have basic config and override it per stack? Something similar to docker-compose definitions:

docker-compose -f docker-compose.yml -f docker-compose.prod.yml up

busy-pizza-73563

07/25/2019, 12:36 PM

Hello, there! I saw that resource adoption (https://github.com/pulumi/pulumi/pull/2893) has just been merged, is there any documentation about that?

glamorous-waitress-51149

07/25/2019, 2:42 PM

anyone see random failures deploying a rds serverless mysql instance? seems to fail, then redeploy and all is well

glamorous-waitress-51149

07/25/2019, 2:42 PM

a bit unnerving

glamorous-waitress-51149

07/25/2019, 2:42 PM

2019-07-25T10:23:20.7618225Z Diagnostics:
2019-07-25T10:23:20.7618650Z   pulumi:pulumi:Stack (profiles-sandbox):
2019-07-25T10:23:20.7620280Z     error: update failed
2019-07-25T10:23:20.7620503Z  
2019-07-25T10:23:20.7620651Z   aws:rds:Cluster (default):
2019-07-25T10:23:20.7620797Z     error: Plan apply failed: 1 error occurred:
2019-07-25T10:23:20.7621447Z         * updating urn:pulumi:sandbox::profiles::aws:rds/cluster:Cluster::default: Failed to modify RDS Cluster (profiles-mysql-aurora): InvalidParameterCombination: You currently can't modify EngineVersion with Aurora Serverless.
2019-07-25T10:23:20.7621983Z         status code: 400, request id: cdfb50b7-cb26-49a5-becc-48fb472e1a22
2019-07-25T10:23:20.7622213Z  
2019-07-25T10:23:20.7622765Z Resources:
2019-07-25T10:23:20.7622890Z     + 69 created
2019-07-25T10:23:20.7623010Z     ~ 5 updated
2019-07-25T10:23:20.7623155Z     2 changes. 49 unchanged
2019-07-25T10:23:20.7623266Z 
2019-07-25T10:23:20.7623390Z Duration: 3m26s

full-dress-10026

07/25/2019, 3:59 PM

What is the recommended way to upgrade your eks node version?

damp-book-35965

07/25/2019, 5:09 PM

Hello..Currently is there a way to list all stacks which use a certain tag ?

future-morning-96441

07/25/2019, 5:37 PM

Did anyone manage to use webpack (especially for treeshaking / reducing dead code) for a Pulumi application? (regarding @pulumi/aws) We have some heavy dependencies, but only use small chunks of them. Pulumi's sereialization results in pretty large deployment packages, due to the dependencies. When I try to use Pulumi in conjunction with webpack and deploy to AWS I hit several issues. Anyone around, who got a similar setup running?

cool-egg-852

07/25/2019, 6:50 PM

When importing, I thought it was supposed to output the actual contents? that’s what it seemed like from https://asciinema.org/a/Ov7s3Xxa0YtXzWwVtleyGYfCw

incalculable-diamond-5088

07/25/2019, 8:03 PM

After upgrading to the latest CLI, each

pulumi up

began showing this diff:

~ kubernetes:policy/v1beta1:PodSecurityPolicy: (update)
    [id=prmetheus-operator-kube-state-metrics]
    [urn=urn:pulumi:staging-account::staging-account::...:account$...:eksCluster$aws:eks/cluster:Cluster$...:eksBasics$...:prometheus$kubernetes:<http://helm.sh/v2:Chart$kubernetes:policy/v1beta1:PodSecurityPolicy::monitoring/prmetheus-operator-kube-state-metrics]|helm.sh/v2:Chart$kubernetes:policy/v1beta1:PodSecurityPolicy::monitoring/prmetheus-operator-kube-state-metrics]>
    [provider=urn:pulumi:staging-account::staging-account::...:account$...:eksCluster$pulumi:providers:kubernetes::k8s::989bfcae-4c24-4ad3-aba9-d1620215508b]
  ~ spec: {
      + hostIPC               : false
      + hostNetwork           : false
      + hostPID               : false
      + privileged            : false
      + readOnlyRootFilesystem: false
    }
~ kubernetes:policy/v1beta1:PodSecurityPolicy: (update)
    [id=prmetheus-operator-prometh-alertmanager]
    [urn=urn:pulumi:staging-account::staging-account::...:account$...a:eksCluster$aws:eks/cluster:Cluster$...:eksBasics$...:prometheus$kubernetes:<http://helm.sh/v2:Chart$kubernetes:policy/v1beta1:PodSecurityPolicy::monitoring/prmetheus-operator-prometh-alertmanager]|helm.sh/v2:Chart$kubernetes:policy/v1beta1:PodSecurityPolicy::monitoring/prmetheus-operator-prometh-alertmanager]>
    [provider=urn:pulumi:staging-account::staging-account::...:account$...:eksCluster$pulumi:providers:kubernetes::k8s::989bfcae-4c24-4ad3-aba9-d1620215508b]
  ~ spec: {
      + hostIPC               : false
      + hostNetwork           : false
      + hostPID               : false
      + privileged            : false
      + readOnlyRootFilesystem: false
    }

Any idea how to get rid of those diffs (and the redundant updates it triggers)?

high-translator-22614

07/25/2019, 8:07 PM

ok, one thing that's never clear: What's the best way to get the current AWS region inside of a

ComponentResource

swift-painter-31084

07/25/2019, 10:55 PM

are there any clever patterns for setting a local environment variable from a Pulumi Output?

early-intern-90238

07/26/2019, 12:27 AM

ok so the way Pulumi handles the creation, and modification of a Kubernetes cluster in GCP is way offbase...

early-intern-90238

07/26/2019, 12:28 AM

For some reason, if you modify a node in the node pool, Pulumi decides to attempt a full replacement of your Cluster (fails at this btw), and creates an entirely new cluster...That's not an idea solution...

early-intern-90238

07/26/2019, 12:28 AM

Instead it should replace just the node that was modified...

early-intern-90238

07/26/2019, 12:28 AM

very confused about this behavior

early-intern-90238

07/26/2019, 12:29 AM

It also did something to completely break my GCP Kubernetes Dashboard as well, I get this error now:

early-intern-90238

07/26/2019, 12:29 AM

I am working with Google support to try to resolve that...

early-intern-90238

07/26/2019, 12:36 AM

That issue in screenshot might be related to Googles outage right now actually