Skip to content

Operator terminates all keeper pods on pod spec changes (no rolling update) #1915

@vstariradev

Description

@vstariradev

Patching a CHK resource's image version triggers puts all chk pods into state Terminating.

Steps to reproduce:

  1. Setup clickhouse operator 0.25.6
  2. Install CHK resource (config shown below)
  3. Patch/Edit image version to 25.3.8.10041.altinitystable
  4. Observe chk pod status

chop.log

Versions:
Altinity-clickhouse-keeper: 25.3.8.10041.altinitystable
Clickhouse-operator: 0.25.6

Slack thread with initial discussion:
https://altinitydbworkspace.slack.com/archives/C02K1MWEK2L/p1770022956770289

kind: ClickHouseKeeperInstallation
metadata:
  annotations:
    meta.helm.sh/release-name: clickhouse-keeper
    meta.helm.sh/release-namespace: monitoring
  creationTimestamp: "2026-02-02T09:27:29Z"
  generation: 5
  labels:
    app.kubernetes.io/managed-by: Helm
    helm.toolkit.fluxcd.io/name: clickhouse-keeper
    helm.toolkit.fluxcd.io/namespace: flux-system
  name: observ01
  namespace: monitoring
  resourceVersion: "1035055898"
  uid: 6f79e099-9ef4-450b-8ce6-bff09727bd51
spec:
  configuration:
    clusters:
    - layout:
        replicasCount: 3
      name: cluster01
  defaults:
    templates:
      podTemplate: clickhouse-keeper-pod-template
      volumeClaimTemplate: clickhouse-keeper-storage-template
  templates:
    podTemplates:
    - metadata:
        labels:
          app: clickhouse-keeper
          app.kubernetes.io/name: clickhouse-keeper
          owner_team: ops-sre
      name: clickhouse-keeper-pod-template
      spec:
        affinity:
          podAntiAffinity:
            requiredDuringSchedulingIgnoredDuringExecution:
            - labelSelector:
                matchExpressions:
                - key: app
                  operator: In
                  values:
                  - clickhouse-keeper
              topologyKey: hw_host
        containers:
        - image: harbor.repo/altinity-clickhouse-keeper:25.3.6.10034.altinitystable
          name: clickhouse-keeper
          resources:
            limits:
              cpu: 500m
              memory: 1Gi
            requests:
              cpu: 200m
              memory: 256Mi
        imagePullSecrets:
        - name: regcred
        nodeSelector:
          tooling: "true"
        securityContext:
          fsGroup: 101
    volumeClaimTemplates:
    - metadata:
        labels:
          owner_team: ops-sre
      name: clickhouse-keeper-storage-template
      reclaimPolicy: Retain
      spec:
        accessModes:
        - ReadWriteOnce
        resources:
          requests:
            storage: 5Gi

Metadata

Metadata

Assignees

No one assigned

    Labels

    KeeperClickHouse Keeper issues

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions