[e2e-testing/tutorial] improve helm-chart

[evegufy]

Description

In order to support e2e-testing and tutorial / local setup, I suggest some improvements to the helm-chart.

Acceptance Criteria

• [BLOCKER FOR INTEGRATION WITH UMBRELLA] solve error when installing with charts that also have a postgresql dependency

I encounter the following error when I install the chart with other that also have a postgresql dependency

Error: UPGRADE FAILED: template: umbrella/charts/discoveryfinder/charts/postgresql/templates/primary/svc.yaml:11:14: executing "umbrella/charts/discoveryfinder/charts/postgresql/templates/primary/svc.yaml" at <include "common.labels.standard" (dict "customLabels" .Values.commonLabels "context" $)>: error calling include: template: umbrella/charts/bpdm/charts/postgres/charts/common/templates/_labels.tpl:6:27: executing "common.labels.standard" at <include "common.names.name" .>: error calling include: template: umbrella/charts/bpdm/charts/postgres/charts/common/templates/_names.tpl:6:18: executing "common.names.name" at <.Chart.Name>: nil pointer evaluating interface {}.Name

Used config:

bpdm:
  bpdm-gate:
    postgres:
      nameOverride: "bpdm-postgresql"
  bpdm-pool:
    postgres:
      nameOverride: "bpdm-postgresql"
  bpdm-bridge-dummy:
    enabled: false
  bpdm-cleaning-service-dummy:
    enabled: false
  postgres:
    nameOverride: "bpdm-postgresql"
    enabled: true

May be an issue connected to the different postgres (the other charts are on bitnami chart version 12.x.x) version and be addressed by #764

• provide images for arm64 cpu arch, example - in addition to amd64 (image doesn't work on newer Macs, this is relevant especially in the tutorial use case)
• set default imagePullPolicy to IfNotPresent
• Suggestion to change the recommended fullnameOverride(s) in the values.yaml to nameOverride(s) - especially in the case when integrating with charts that have the same dependency, unique labels (name) are important

Additional Information

evaluation done on chart version 4.0.1

cc: @jzbmw

[almadigabor]

Hey @evegufy, good suggestions there, even trying to install the chart directly does not really work for me as most of the pods do not reach the Ready state and restart constantly. This also results in pulling the images on each restart as you described the lack of default IfNotPresent pull policy.

helm install subchart tractusx-dev/bpdm -n bpdm --create-namespace --version 4.0.1

# kubectl get pods -n bpdm -o wide
NAME                                                    READY   STATUS    RESTARTS        AGE   IP            NODE                   NOMINATED NODE   READINESS GATES
subchart-postgres-0                                     1/1     Running   0               15m   10.42.0.169   lima-rancher-desktop   <none>           <none>
subchart-bpdm-pool-66f67f4499-lgns9                     0/1     Running   1 (3m12s ago)   15m   10.42.0.167   lima-rancher-desktop   <none>           <none>
subchart-bpdm-gate-fb68ff667-8ftgq                      0/1     Running   2 (108s ago)    15m   10.42.0.164   lima-rancher-desktop   <none>           <none>
subchart-bpdm-orchestrator-65675b7977-7kr4f             0/1     Running   2 (93s ago)     15m   10.42.0.165   lima-rancher-desktop   <none>           <none>
subchart-bpdm-bridge-dummy-58b4b9c89-k2758              0/1     Running   2 (78s ago)     15m   10.42.0.163   lima-rancher-desktop   <none>           <none>
subchart-bpdm-cleaning-service-dummy-7964bdbd6f-d27lp   0/1     Running   2 (48s ago)     15m   10.42.0.166   lima-rancher-desktop   <none>           <none>

An example on how to implement multiple docker image architectures can be found here.

[evegufy]

Hey @evegufy, good suggestions there, even trying to install the chart directly does not really work for me as most of the pods do not reach the Ready state and restart constantly. This also results in pulling the images on each restart as you described the lack of default IfNotPresent pull policy.

helm install subchart tractusx-dev/bpdm -n bpdm --create-namespace --version 4.0.1

# kubectl get pods -n bpdm -o wide
NAME                                                    READY   STATUS    RESTARTS        AGE   IP            NODE                   NOMINATED NODE   READINESS GATES
subchart-postgres-0                                     1/1     Running   0               15m   10.42.0.169   lima-rancher-desktop   <none>           <none>
subchart-bpdm-pool-66f67f4499-lgns9                     0/1     Running   1 (3m12s ago)   15m   10.42.0.167   lima-rancher-desktop   <none>           <none>
subchart-bpdm-gate-fb68ff667-8ftgq                      0/1     Running   2 (108s ago)    15m   10.42.0.164   lima-rancher-desktop   <none>           <none>
subchart-bpdm-orchestrator-65675b7977-7kr4f             0/1     Running   2 (93s ago)     15m   10.42.0.165   lima-rancher-desktop   <none>           <none>
subchart-bpdm-bridge-dummy-58b4b9c89-k2758              0/1     Running   2 (78s ago)     15m   10.42.0.163   lima-rancher-desktop   <none>           <none>
subchart-bpdm-cleaning-service-dummy-7964bdbd6f-d27lp   0/1     Running   2 (48s ago)     15m   10.42.0.166   lima-rancher-desktop   <none>           <none>

An example on how to implement multiple docker image architectures can be found here.

maybe related to the finding from @FaGru3n #786

[nicoprow]

I provided a fix for this issue here #800

Seems chart testing did not catch that one. The fixed chart now installs without any other configuration.

Regarding the other suggestions, I think they are good but I will provide such improvements for the next release.

[nicoprow]

Nobody working on that now and I don't see the capacity for it in the near future. I close this issue for now and we can reevaluate this workflow feature at a later time