Not able to execute the script wait-for-cluster.sh

[taruntmr]

Hi,

I am facing issue when consuming the module private-cluster. I get the following error:

[0m�[0mmodule.k8s-private-cluster.null_resource.wait_for_cluster[0] (local-exec): /bin/sh: terraform-google-kubernetes-engine/modules/private-cluster/scripts/wait-for-cluster.sh: not found

I am using the hashicorp/terraform:0.12.16 image.

I connected to the running container and tried executing the file manually but I got the same error.

I validated the permission on the file, they are +x.

Let me know if you would want more details.

Thanks,
Tarun

[rohitgabrieleit]

Facing the same issue, any thoughts on how to resolve this?

[morgante]

I'm not really sure what could be causing this truthfully. It looks like the script is there, yet it's somehow not being executed?

Could you try directly executing the script (outside Terraform)?

[rohitgabrieleit]

I am using Terraform cloud, so that's not an option

[rohitgabrieleit]

Is this a consequence of
gcloud executable requirement, removal of jq #401 . #401

[rohitgabrieleit]

Error: Error running command '.terraform/modules/gke/terraform-google-modules-terraform-google-kubernetes-engine-1d4c858/scripts/wait-for-cluster.sh friendly-path-269502 simple-regional-cluster-pih-sandbox': exit status 127.
Output: Waiting for cluster simple-regional-cluster-pih-sandbox in project friendly-path-269502 to reconcile...

.terraform/modules/gke/terraform-google-modules-terraform-google-kubernetes-engine-1d4c858/scripts/wait-for-cluster.sh: line 28: gcloud: command not found

[mouglou]

I get the same issue... The cluster is good, but the job is marked as "passed with warnings" in my Gitlab CICD. So not really cool ;)

I try few thinkgs but don't find a solution for now

I add an "ls" and try to execute the script manually without success.

$ ls -al .terraform/modules/gke/terraform-google-modules-terraform-google-kubernetes-engine-1d4c858/modules/private-cluster/scripts/
57 total 20
58 drwxrwxr-x    2 root     root          4096 Feb 19 16:23 .
59 drwxrwxr-x    3 root     root          4096 Feb 19 16:23 ..
60 -rwxrwxr-x    1 root     root          1585 Feb 19 16:23 delete-default-resource.sh
61 -rwxrwxr-x    1 root     root          1573 Feb 19 16:23 kubectl_wrapper.sh
62 -rwxrwxr-x    1 root     root          1236 Feb 19 16:23 wait-for-cluster.sh
63 $ .terraform/modules/gke/terraform-google-modules-terraform-google-kubernetes-engine-1d4c858/modules/private-cluster/scripts/wait-for-cluster.sh coretech-ludia luna-gke-srvr-01
64 /bin/sh: eval: line 97: .terraform/modules/gke/terraform-google-modules-terraform-google-kubernetes-engine-1d4c858/modules/private-cluster/scripts/wait-for-cluster.sh: not found

[morgante]

I suspect the issue might be running in an environment where bash is not available. We should change the script to specify /bin/sh instead of /bin/bash: https://github.com/terraform-google-modules/terraform-google-kubernetes-engine/blob/master/autogen/main/scripts/wait-for-cluster.sh#L1

[mouglou]

Just to let you know, I try to change my gitlab-runner from "docker" to "shell" and with that, the script is working fine. So maybe the "not found" is coming from the "gcloud" package that is missing in the "terraform:light" docker image.

[dexter0195]

I suspect the issue might be running in an environment where bash is not available. We should change the script to specify /bin/sh instead of /bin/bash: https://github.com/terraform-google-modules/terraform-google-kubernetes-engine/blob/master/autogen/main/scripts/wait-for-cluster.sh#L1

My issue was exactly that, I am running terraform in a docker container and bash was not available.
Installing bash solved the issue

[mbrowne]

For me, this error went away when I copied all the files in this module to a local directory, source = "./beta-private-cluster" (instead of source = "terraform-google-modules/kubernetes-engine/google//modules/beta-private-cluster").

I'm also using terragrunt, so that might be a complicating factor.

Earlier I had tried setting skip_provisioners = false, which supposedly should have prevented it from running the shell scripts (including wait-for-cluster), but it was still complaining that wait-for-cluster.sh wasn't found.

[llnformer]

any idea how to overcome the issue of not having the gcloud command when running in Terraform Cloud environment?

[bharathkkb]

Hi @llnformer
skip_provisioners = false should now let you skip wait-for-cluster.sh, there was a bug that was fixed by #669

Additionally with the 12.0 release, if you set the GCLOUD_TF_DOWNLOAD env var, this module will attempt to download and install gcloud. More details here https://github.com/terraform-google-modules/terraform-google-kubernetes-engine#gcloud

Closing this issue out now, feel free to reopen if needed.

[jai]

skip_provisioners = false should now let you skip wait-for-cluster.sh, there was a bug that was fixed by #669

@bharathkkb you mean skip_provisioners = true yeah? Or am I misunderstanding?