1.1.0 breaks ct install workflow about chart-testing-action

[davidkarlsen]

Chart testing runs as a container, so I wonder about the endpoint localhost:8080. Unless localhost is actually both the host running docker as the container and hosting the k8s api endpoint it won't work.

One thing you can experiment with though is to keep the updated action version, but downgrade the ct image: https://github.com/helm/chart-testing-action/blob/master/README.md#L21 - to see if that changes anything.

[steven-sheehy]

localhost:8080 is the default in ~/.kube/config up until 1.18. So this means [email protected] is not using the custom kubeconfig I'm passing via --kubeconfig and is using the default, which of course won't work.

I tested it with helm/[email protected] and image: quay.io/helmpack/chart-testing:v3.0.0 and it fails with the below error. I tested it with helm/[email protected] and image: quay.io/helmpack/chart-testing:v3.1.1 and it succeeds.

Run helm/[email protected]
  with:
    command: install
    config: .github/ct-install.yaml
    image: quay.io/helmpack/chart-testing:v3.0.0
    kubeconfig: /tmp/output/kubeconfig-v1.17.4-k3s1.yaml
  env:
    KUBECONFIG: /tmp/output/kubeconfig-v1.17.4-k3s1.yaml
Running ct container...
48e699207b18c61aebbeec1df6857443cb2b938e34274ba1fd3931245772841a

Running 'ct install'...
Installing charts...
Using config file:  /etc/ct/ct.yaml
Version increment checking disabled.
------------------------------------------------------------------------------------------------------------------------
 Configuration
------------------------------------------------------------------------------------------------------------------------
Remote: origin
TargetBranch: master
Since: HEAD
BuildId: 
LintConf: /etc/ct/lintconf.yaml
ChartYamlSchema: /etc/ct/chart_schema.yaml
ValidateMaintainers: false
ValidateChartSchema: false
ValidateYaml: false
CheckVersionIncrement: false
ProcessAllCharts: false
Charts: [charts/hedera-mirror]
ChartRepos: [bitnami=https://charts.bitnami.com/bitnami fluxcd=https://charts.fluxcd.io loki=https://grafana.github.io/loki/charts prometheus=https://prometheus-community.github.io/helm-charts traefik=https://helm.traefik.io/traefik]
ChartDirs: [charts]
ExcludedCharts: []
HelmExtraArgs: 
HelmRepoExtraArgs: []
Debug: false
Upgrade: false
SkipMissingValues: false
Namespace: 
ReleaseLabel: app.kubernetes.io/instance
------------------------------------------------------------------------------------------------------------------------

------------------------------------------------------------------------------------------------------------------------
 Charts to be processed:
------------------------------------------------------------------------------------------------------------------------
 hedera-mirror => (version: "0.7.0-alpha1", path: "charts/hedera-mirror")
------------------------------------------------------------------------------------------------------------------------

"bitnami" has been added to your repositories
"fluxcd" has been added to your repositories
"loki" has been added to your repositories
"prometheus" has been added to your repositories
"traefik" has been added to your repositories
Hang tight while we grab the latest from your chart repositories...
...Successfully got an update from the "traefik" chart repository
...Successfully got an update from the "loki" chart repository
...Successfully got an update from the "fluxcd" chart repository
...Successfully got an update from the "prometheus" chart repository
...Successfully got an update from the "bitnami" chart repository
Update Complete. ⎈Happy Helming!⎈
Saving 5 charts
Downloading postgresql-ha from repo https://charts.bitnami.com/bitnami
Downloading redis from repo https://charts.bitnami.com/bitnami
Deleting outdated charts
Installing chart 'hedera-mirror => (version: "0.7.0-alpha1", path: "charts/hedera-mirror")'...

Installing chart with values file 'charts/hedera-mirror/ci/default-values.yaml'...

Creating namespace 'hedera-mirror-nc1x0zahxl'...
error: Missing or incomplete configuration info.  Please point to an existing, complete config file:

  1. Via the command-line flag --kubeconfig
  2. Via the KUBECONFIG environment variable
  3. In your home directory as ~/.kube/config

To view or setup config directly use the 'config' command.
========================================================================================================================
........................................................................................................................
==> Events of namespace hedera-mirror-nc1x0zahxl
........................................................................................................................
error: Missing or incomplete configuration info.  Please point to an existing, complete config file:

  1. Via the command-line flag --kubeconfig
  2. Via the KUBECONFIG environment variable
  3. In your home directory as ~/.kube/config

To view or setup config directly use the 'config' command.
Error printing details: Error waiting for process: exit status 1
Error printing logs: Error running process: exit status 1
Deleting release 'hedera-mirror-nc1x0zahxl'...
Error: Kubernetes cluster unreachable
Error deleting Helm release: Error waiting for process: exit status 1

So it seems the problem is located in chart-testing-action. Most likely due to the removal of the docker_exec sh -c 'mkdir -p /root/.kube' in this commit?

[davidkarlsen]

Hm, the copy should copy the whole directory after that change, and not only the kubeconfig file - which was the reason for the change in #46 - how does the contents of the directory which "${{ steps.k8s.outputs.kubeconfig }}" reside in look like?

[davidkarlsen]

@steven-sheehy ping? any update - or did you get it to work?

[steven-sheehy]

No, it still doesn't work. I worked around it by downgrading to 1.0.0. Here's the ls output you requested:

Run ls -la "/tmp/output/kubeconfig-v1.17.4-k3s1.yaml"
-rw-rw-rw- 1 root root 1052 Oct 23 23:45 /tmp/output/kubeconfig-v1.17.4-k3s1.yaml

[davidkarlsen]

This should solve itself with v2.0.0 as it simply installs and makes ct avail from PATH.

[davidkarlsen]

Closing this as it was fixed in v2.x