pulumi-community #general

Join the conversation Join Slack

general

tall-librarian-49374

06/18/2019, 6:53 AM

Great new flow for getting started! My minor remarks below.

👍 1

faint-motherboard-95438

06/18/2019, 2:27 PM

Hi there, I think I noticed a bug, or maybe I’m doing something wrong. I’d like your advice before opening an issue. When I update a

Deployment

configuration, I noticed it creates a new replicaset and does not delete the old one (or maybe it should update it, I don’t know). So it actually kills all the old pods in the old replicaset and creates the new one as expected, but now I have a stale useless replicasets. Is this something I misconfigured somewhere or is this a bug ?

replicaset.apps/context-clusters-pvmtest-mqtt-5979fbcdbc                       2         2         2       2m13s
replicaset.apps/context-clusters-pvmtest-mqtt-client-68fd4d4445                0         0         0       23m
replicaset.apps/context-clusters-pvmtest-mqtt-client-6b8bdfc4b7                2         2         2       2m13s
replicaset.apps/context-clusters-pvmtest-mqtt-f5bd9548f                        0         0         0       25m
replicaset.apps/context-clusters-pvmtest-portal-567948dccd                     2         2         2       2m13s
replicaset.apps/context-clusters-pvmtest-portal-6c5968cfcd                     0         0         0       26m

bulky-businessperson-73745

06/18/2019, 3:53 PM

hey all. i’m excited about pulumi and giving it a spin. the link for “pulumi community edition” on the pricing page seems to take me to the home page. is this intended? i was hoping to see what was included with the community edition.

elegant-crayon-4967

06/18/2019, 4:20 PM

Does anyone have a working example of using an external user_data script w/ variables defined in

index.ts

for aws & typescript?

limited-rainbow-51650

06/18/2019, 4:20 PM

How can I import existing resources into the Pulumi state? Stated differently, what is the Pulumi equivalent to

terraform state import

(https://www.terraform.io/docs/import/index.html)

proud-alarm-92546

06/18/2019, 5:09 PM

having trouble doing a password reset via pulumi.com. get a

not found: verification code "code" not found

(which entirely amusing in it's repetition..

elegant-crayon-4967

06/18/2019, 5:47 PM

to build on my previous question...looking for the user_data to be a powershell script

handsome-actor-1155

06/18/2019, 6:15 PM

Hello, I have a question regarding deployments of pods to to k8s via Pulumi. I have a deployment that builds a docker image, uploads to ECR, and then deploys a replica set to k8s. When I update the docker image, I would expect a redeploy of the replica set in k8s but I do not see that happening.

Updating (dev):

     Type                             Name                                       Status      Info
     pulumi:pulumi:Stack              stream-demo-sensor-reading-aggregator-dev              
 >   ├─ pulumi:pulumi:StackReference  brianbrifri/stream-demo-services/dev       read        
 >   ├─ pulumi:pulumi:StackReference  brianbrifri/stream-demo-infra/dev          read        
     ├─ docker:image:Image            sensorReadingAggregator                                
 ~   └─ kubernetes:apps:ReplicaSet    stream-demo-sensor-reading-aggregator      updated     [diff: ~spec]

Is my output but the replica set in k8s is unchanged. Is this expected behavior?

proud-alarm-92546

06/18/2019, 8:04 PM

anyone available to discuss local environment setup? running into an interesting set of failures on

pulumi/pulumi

first time running

make

on a stock master,

golangci-lint run

complains about a bunch of things that don't seem to be erroring in the travis builds. makes me wonder if there's

golangci-link

version issue, or if there's a config somewhere that's not in the repo, or......

cmd/pulumi-language-nodejs/main.go:87:56: `langauge` is a misspelling of `language` (misspell)
	cmdutil.InitTracing("pulumi-language-nodejs", "pulumi-langauge-nodejs", tracing)

(if I fix that one, it moves on to many others) opened an issue - https://github.com/pulumi/pulumi/issues/2841

full-dress-10026

06/18/2019, 9:36 PM

Are there any examples of using an Application Gateway with App Service in Azure?

busy-air-16359

06/19/2019, 12:29 PM

-.txt

busy-air-16359

06/19/2019, 12:29 PM

I’m getting weird connection refused errors out of the sudden, has anybody seen something like this?

busy-air-16359

06/19/2019, 12:29 PM

followed by a stacktrace from pulumi

thankful-lamp-48382

06/19/2019, 1:59 PM

Hello guys! Pulumi looks really interesting! I need some simpler example to get started, something like create Droplet, upload some node.js app to it and execute npm install & npm run start on that Droplet, can also be for Linode

thankful-lamp-48382

06/19/2019, 1:59 PM

Do you know any tutorial like this?

thankful-lamp-48382

06/19/2019, 2:22 PM

One more thing that bothers me: I am working on multiplayer game and I would have initially 4 instances (Droplets) with game rooms, but I would like to dynamically add new Droplet when these initial 4 are full. Can I achieve it somehow with Pulumi?

thankful-lamp-48382

06/19/2019, 2:30 PM

I mean I would detect myself that rooms are full, then would call some web hook or something that would call Pulumi? Is something like this possible?

boundless-monkey-50243

06/19/2019, 3:18 PM

Is there a way, akin to

terraform taint

, that I can tell Pulumi “that resource must be replaced”? (I don’t think I can manually remove it; it’s an AWS API Gateway deployment and those are archived rather than deleted.)

full-dress-10026

06/19/2019, 4:17 PM

The sample here: https://pulumi.io/reference/pkg/nodejs/pulumi/azure/compute/#example-usage-with-managed-disks-recommended on this line:

loadBalancerInboundNatRulesIds: [pulumi.all(lbnatpool.map(v => v.id)).apply(id => id.map(v => v)[1])],

does not work. You cannot

map

NatPool

broad-dog-22463

06/19/2019, 4:18 PM

@full-dress-10026 please can you open a bug report in pulumi-azure repo?

bored-river-53178

06/19/2019, 4:40 PM

would it be possible to reopen https://github.com/pulumi/pulumi-awsx/issues/314 as it is still not fixed?

early-musician-41645

06/19/2019, 5:53 PM

I saw this same error earlier here: https://pulumi-community.slack.com/archives/C84L4E3N1/p1558201564057000?thread_ts=1558193224.055800&cid=C84L4E3N1 However, I am running

pulumi up

and the error outputs:

error: Plan apply failed: Program run without the `pulumi` CLI; this may not be what you want (enable PULUMI_TEST_MODE to disable this error)

This started happening after I defined a new function as part of my custom resource I'm using Pulumi

v0.17.16

Any idea on a fix?

full-dress-10026

06/19/2019, 6:01 PM

When using a

azure.compute.ScaleSet

to provision VMs, can you have the scale set assign public IPs to the created VMs so you can ssh onto them?

early-musician-41645

06/19/2019, 6:56 PM

I'm running into difficulties with async functions in typescript with Pulumi CLI

19 async function getServiceAccountAuth() {
 20   const secret = await aws.secretsmanager.getSecretVersion({
 21     secretId: "some-secret-name",
 22   });
 23   const secretString = secret.secretString
 24   console.log("SECRETSTRING: "+secretString);
 25
 26   let username: string = JSON.parse(secretString)["username"];
 27   let password: string = JSON.parse(secretString)["password"];
 28
 29   console.log("username: " + username);
 30
 31   return "Basic " + new Buffer(username + ":" + password).toString("base64");
 32 }

When I run that function I get this:

error: Plan apply failed: Program run without the `pulumi` CLI; this may not be what you want (enable PULUMI_TEST_MODE to disable this error)

What am I doing wrong?

full-dress-10026

06/19/2019, 7:01 PM

If I change the

osProfile

of a

ScaleSet

I get this outputted from `pulumi up`:

azure:compute:ScaleSet  azureDemo-ss2             updated     [diff: ~osProfile]

However, the actual instances are not restarted. When changing

customData

, I would expect instances to restart to apply the new customData.

full-dress-10026

06/19/2019, 9:04 PM

Is there a way to have stack outputs that are outputs but not shown at the end of

pulumi up

? Use case would be outputting a kubeconfig.

➕ 1

full-dress-10026

06/19/2019, 11:11 PM

Is there an easy path to get something like this https://www.terraform.io/docs/providers/datadog/index.html to work with Pulumi?

late-lawyer-23405

06/19/2019, 11:27 PM

Main web site says to request languages here: I'm interested in Clojure, ClojureScript, and Kotlin

colossal-room-15708

06/19/2019, 11:30 PM

Just an idea to the channel owners... should we maybe start splitting providers into their own channels? #general is getting really noisy and lots of messages drown in the noise (@white-balloon-205?)

early-musician-41645

06/19/2019, 11:55 PM

Is there a way to pass a value to the

async delete

method of a dynamic resource provider?

Title

early-musician-41645

06/19/2019, 11:55 PM

Is there a way to pass a value to the

async delete

method of a dynamic resource provider?

async delete(id, props)

method requires some auth arguments to form a REST

DELETE

request. For the

async create(inputs)

I just added the auth data into the

inputs

and had no trouble. How can I make those available to

delete

white-balloon-205

06/20/2019, 12:01 AM

You can reference variables from the enclosing scope that contain these credentials - getting the values out of globals (perhaps from Pulumi config), or from parameters passed in to the Provider implementation somehow.

early-musician-41645

06/20/2019, 12:01 AM

It's more complicated - the auth is pulled from aws secretsmanager

white-balloon-205

06/20/2019, 12:01 AM

Putting these in the

inputs

probably isn't right here - there won't be any inputs during a

delete

early-musician-41645

06/20/2019, 12:02 AM

right, so how do I pass it to the provider?

white-balloon-205

06/20/2019, 12:03 AM

It's more complicated - the auth is pulled from aws secretsmanager

That should still be possible. Is it pulled from secrets manager at deployment time via

aws.secretsmanager.getSecretVersion

early-musician-41645

06/20/2019, 12:04 AM

I spent a couple days trying to get that method to work and eventually just used the

aws-sdk

instead. So the secret is pulled like this:

28 var secretsmanager = new AWS.SecretsManager();
 29 secretsmanager.getSecretValue({ SecretId: "eshamay-test" }, function(err: any, data: any) {
 30   if (err) {
 31     console.log("ERROR getting secret for service account", err);
 32     throw err;
 33   } else {
 34     const savedSearch = new splunk.SavedSearch("foo-search", {
 35       name: "foo-eshamay-test-1",
 36       description: "eshamay's test search",
 37       search: 'search index=online_prod source="*10ay*" earliest=-5m | stats count as Total',
 38       splunkAuthSecret: data.SecretString,
 39     });
 40   }
 41 });
 42

the SavedSearch is my

pulumi.dynamic.Resource

so the resource is created in the callback

I worked around this by doing this:

81 export class SavedSearch extends pulumi.dynamic.Resource {
 82   public readonly name: pulumi.Output<string>;
 83
 84   private static provider = new SavedSearchProvider();
 85
 86   constructor(name: string, props: SavedSearchInput, opts?: pulumi.CustomResourceOptions) {
 87     super(SavedSearch.provider, name, props, opts);
 88
 89     SavedSearch.provider.setAuth(props.splunkAuthSecret);
 90   }
 91 }

The creds are still in the

inputs

, but then I added a setter to the provider to create an instance variable that I can use in the

delete

method

Not ideal, but it works

white-balloon-205

06/20/2019, 12:16 AM

Yeah - something along those lines is likely going to be the best option. Writing a short example of a possibly slightly "nicer" interface now that I'll share.

Actually - going back to the original question. If you do want to pass this data through the resource properties, your

delete

implementation is passed the old

props

for the resource, so any credentials provided at creation time will be available on the

props

. Was that not working for you? Note that the Pulumi program does not run at all during a

destroy

- so there is no way to pass new values in during the deletes for a destroy - the only data you have available to you is either: 1. The old inputs from when the resource was created/updated. 2. The serialized implementation of the

delete

body, which itself may have serialized in some state captured during that last update.

View count: 1