While I expect the environment variable in the Deployment to be set to

<http://api-vvf3g61k>

(or similar)

I also tried to replace the

srv_name1

to

s[0][0].metadata.name

instead. Doing this will set the environment variable to some other

Output

object, not the metadata name I’m looking for.

Hmm

Also wrapping the whole deployment in the

Output.all().apply(lambda …)

is not very readable.

Maybe I need something like

Output.all([central_api_deployment.status])

instead of

Output.all([central_api_deployment])

?

maybe

you’ll figure it out, just keep playing with it. once it clicks you’ll get it

Thanks again @sparse-state-34229. I really appreciate your help.

np!

have anybody encountered this errors when using kube_config generated from gke cluster? configured Kubernetes cluster is unreachable: failed to parse kubeconfig data in `kubernetesconfigkubeconfig`; this must be a YAML literal string and not a filename or path - yaml: line 2: mapping values are not allowed in this context

code:

cluster_info = pulumi.Output.all(cluster.name, cluster.endpoint, cluster.master_auth)
kube_config = cluster_info.apply(
    lambda info: """apiVersion: v1
clusters:
- cluster:
    certificate-authority-data: {0}
    server: https://{1}
  name: {2}
contexts:
- context:
    cluster: {2}
    user: {2}
  name: {2}
current-context: {2}
kind: Config
preferences: {{}}
users:
- name: {2}
  user:
    auth-provider:
      config:
        cmd-args: config config-helper --format=json
        cmd-path: gcloud
        expiry-key: '{{.credential.token_expiry}}'
        token-key: '{{.credential.access_token}}'
      name: gcp
""".format(cluster_info[2]['clusterCaCertificate'], cluster_info[1], '{0}_{1}_{2}'.format(
        config.customer_name, config.env_name, config.gcp_region)))

it is taken from pulumi examples on github https://github.com/pulumi/examples/blob/master/gcp-py-gke/__main__.py

the issue was networking not the example code.

Good Morning. I create resources in aws using python. Works great 🙂 Now I would like to render an inventory template for use in ansible. To that end I need the IP addresses of my servers. They are available in

server.public_ip

, however when I print that (I actually render it in a jinja2 template), its an pulumi.output.Output Object. I read in the documentation that this is because of the asynchroneous nature of the object. I tried to convert that Output object to a string using

pulumi.Output.concat()

, however that did not do the trick. I guess I would need to somehow define the dependency bevor using that, however I am a bit lost on how to do all that. Can someone explain, or has an example on how to get IP addresses from servers for further processing?

are you using ansible within pulumi?

can you export them and use

pulumi stack output -j

to process?

Well, my current model (or idea) is that I run pulumi to create the infrastructure, and pulumi also creates an inventory file (containing not only IP addresses, but also some variables for ansible, which are defined in the jinja2 template). So after pulumi terminates, I call ansible to do its work. So not really within pulumi.

I am not sure what

pulumi stack output -j

does, but it does not really sound like I could render my custom inventory template.

export the IP addresses, write a shell script to run that command and create the inventory file

I see. That sure would work (I am calling pulumi from within a shell script anyway to deal with mfa token auth and python virtualenv). But since pulumi has a way of eventually getting the IP address from the Output object, I should be able to do the same thing. I am curious how that would be done.

And thank you for the hint with stack output 🙂

🙂 well if you run out of time figuring out .apply you have another option!

Exactly - in the end its about getting it done.

yup! are you calling .apply when writing to a file? or before that

Nope. I shall try that.

since the apply is a deferred function, i’ve found it only works on the final method using the value

ok. "final" as in the last method in the code that acts uppon the value? I was wondering if there is some sort of

.wait()

method that would block until all tasks related to the object have finished_

?

re: final - i belive so, yes