creating crds with the helm chart fails with configmap too long

[amithalle]

When trying to apply the helm chart, we are getting the following error:
Failed sync attempt to ####: one or more objects failed to apply, reason: ConfigMap "altinity-clickhouse-operator-crds" is invalid: metadata.annotations: Too long: may not be more than 262144 bytes (retried 5 times).

helm chart version: 0.25.6

The issue was resolved the issue locally by splitting the configmap into 2 parts:

Values.crdHook.enabled }}
apiVersion: v1
kind: ConfigMap
metadata:
  name: {{ include "altinity-clickhouse-operator.fullname" . }}-crds-part1
  namespace: {{ include "altinity-clickhouse-operator.namespace" . }}
  labels:
    {{- include "altinity-clickhouse-operator.labels" . | nindent 4 }}
    app.kubernetes.io/component: crd-install-hook
  annotations:
    "helm.sh/hook": pre-install,pre-upgrade
    "helm.sh/hook-weight": "-7"
    "helm.sh/hook-delete-policy": before-hook-creation,hook-succeeded
data:
  clickhouseinstallationtemplates.yaml: |
{{ .Files.Get "crds/CustomResourceDefinition-clickhouseinstallationtemplates.clickhouse.altinity.com.yaml" | indent 4 }}
  clickhousekeeperinstallations.yaml: |
{{ .Files.Get "crds/CustomResourceDefinition-clickhousekeeperinstallations.clickhouse-keeper.altinity.com.yaml" | indent 4 }}
  clickhouseoperatorconfigurations.yaml: |
{{ .Files.Get "crds/CustomResourceDefinition-clickhouseoperatorconfigurations.clickhouse.altinity.com.yaml" | indent 4 }}
---
apiVersion: v1
kind: ConfigMap
metadata:
  name: {{ include "altinity-clickhouse-operator.fullname" . }}-crds-part2
  namespace: {{ include "altinity-clickhouse-operator.namespace" . }}
  labels:
    {{- include "altinity-clickhouse-operator.labels" . | nindent 4 }}
    app.kubernetes.io/component: crd-install-hook
  annotations:
    "helm.sh/hook": pre-install,pre-upgrade
    "helm.sh/hook-weight": "-7"
    "helm.sh/hook-delete-policy": before-hook-creation,hook-succeeded
data:
  clickhouseinstallations.yaml: |
{{ .Files.Get "crds/CustomResourceDefinition-clickhouseinstallations.clickhouse.altinity.com.yaml" | indent 4 }}

{{- end }}

then fix in the job's pod spec:

    spec:
      serviceAccountName: {{ include "altinity-clickhouse-operator.fullname" . }}-crd-install
      restartPolicy: OnFailure
      {{- with .Values.crdHook.nodeSelector }}
      nodeSelector:
        {{- toYaml . | nindent 8 }}
      {{- end }}
      {{- with .Values.crdHook.affinity }}
      affinity:
        {{- toYaml . | nindent 8 }}
      {{- end }}
      {{- with .Values.crdHook.tolerations }}
      tolerations:
        {{- toYaml . | nindent 8 }}
      {{- end }}
      containers:
      - name: crd-install
        image: "{{ ( tpl $.Values.crdHook.image.repository $ ) }}:{{ ( tpl $.Values.crdHook.image.tag $ ) }}"
        imagePullPolicy: {{ .Values.crdHook.image.pullPolicy | default "IfNotPresent" }}
        command:
        - /bin/sh
        - -c
        - |
          set -e
          echo "Installing/Updating ClickHouse Operator CRDs..."
          for crd_file in /crds/*/*.yaml; do
            echo "Applying $(basename $crd_file)..."
            kubectl apply --server-side=true --force-conflicts -f "$crd_file"
          done
          echo "CRD installation completed successfully"
        volumeMounts:
        - name: crds-part1
          mountPath: /crds/part1
          readOnly: true
        - name: crds-part2
          mountPath: /crds/part2
          readOnly: true
        {{- with .Values.crdHook.resources }}
        resources:
          {{- toYaml . | nindent 10 }}
        {{- end }}
      volumes:
      - name: crds-part1
        configMap:
          name: {{ include "altinity-clickhouse-operator.fullname" . }}-crds-part1
      - name: crds-part2
        configMap:
          name: {{ include "altinity-clickhouse-operator.fullname" . }}-crds-part2
        mountPath: /crds
        readOnly: true

[Slach]

Thanks for reporting!
What is your kubernetes installation? Which kubernetes version?

Which cloud priovider EKS, GKE, AKS?

looks like your metadata.annotations insideconfigmap contains something else before applying helm chart

[andrei-dascalu]

@amithalle I believe this could be solved by using server side apply (though I believe this involves rendering the chart and then passing it to Kubectl with server side apply flag).