# Upgrade to Everest 1.8.0 fails with context canceled due to hanging pre-upgrade hook **URL:** \ **Category:** Percona Everest\ **Created:** [August 8, 2025, 8:27am UTC](https://forums.percona.com/t/upgrade-to-everest-1-8-0-fails-with-context-canceled-due-to-hanging-pre-upgrade-hook/39026 "2025-08-08T08:27:02Z")\ **Posts on this page:** 4\ **Page:** 1
**Author:** ![pstekunov](https://sea1.discourse-cdn.com/flex019/user_avatar/forums.percona.com/pstekunov/32/20328_2.png) [@pstekunov](https://forums.percona.com/u/pstekunov)\ **Post date:** [August 8, 2025, 8:27am UTC](https://forums.percona.com/t/upgrade-to-everest-1-8-0-fails-with-context-canceled-due-to-hanging-pre-upgrade-hook/39026/1 "2025-08-08T08:27:02Z")
Sure! Here’s the bug report text translated into clear and professional English: * * * **Title:** Upgrade to Everest 1.8.0 fails with `context canceled` due to hanging pre-upgrade hook * * * ### 📌 Problem Description Attempting to upgrade the `everest-core` release to version **1.8.0** fails with the following issues: - Helm upgrade fails with: - The `everest-helm-pre-upgrade-hook-` Job remains in the `active` state and never completes. - The job logs contain an authentication error when trying to interact with the Kubernetes API: * * * ### 🔁 Steps to Reproduce 1. Run the upgrade command: 2. Helm attempts to run the `pre-upgrade` hook job. 3. The job remains active and never completes, eventually resulting in: * * * ### 🔍 Details - The pre-upgrade hook job is created but does not finish. - Job logs: * * * ### 🛠 Environment - **Chart:** `percona/everest` v1.8.0 - **Kubernetes:** v1.30.2 (OIDC + RBAC enabled) - **Helm:** v3.x - **Namespace:** `everest-system` * * * ### 💡 Root Cause Hypothesis The hook job is likely trying to query the Kubernetes API or the Everest API, but fails due to: - Missing or misconfigured `ServiceAccount` or insufficient RBAC permissions. - Cluster requires external authentication (e.g., OIDC), and the in-cluster job does not have access to proper credentials. * * * ### ✅ Expected Behavior The pre-upgrade hook should either: - Complete successfully if permissions are sufficient, **or** - Fail clearly with a descriptive error — not hang indefinitely. The Helm upgrade should not fail with a timeout and “context canceled” if version discovery fails. * * * ### 📣 Recommendations - Improve error handling in the pre-upgrade hook. - Document required permissions and service account roles for upgrade hooks. - Add fallback behavior when version check fails due to auth issues. * * * Let me know if you want this in Markdown format or ready for GitHub/Percona submission. ---
**Author:** ![pstekunov](https://sea1.discourse-cdn.com/flex019/user_avatar/forums.percona.com/pstekunov/32/20328_2.png) [@pstekunov](https://forums.percona.com/u/pstekunov)\ **Post date:** [August 8, 2025, 8:37am UTC](https://forums.percona.com/t/upgrade-to-everest-1-8-0-fails-with-context-canceled-due-to-hanging-pre-upgrade-hook/39026/2 "2025-08-08T08:37:56Z")
I added permissions for service acount: apiVersion: [rbac.authorization.k8s.io/v1](http://rbac.authorization.k8s.io/v1) kind: ClusterRoleBinding metadata: name: everest-helm-pre-upgrade-hook-binding labels: [app.kubernetes.io/component:](http://app.kubernetes.io/component:) upgrade-hook [app.kubernetes.io/name:](http://app.kubernetes.io/name:) everest subjects: - kind: ServiceAccount name: everest-helm-pre-upgrade-hook namespace: everest-system roleRef: kind: ClusterRole name: everest-pre-upgrade-hook-role apiGroup: [rbac.authorization.k8s.io](http://rbac.authorization.k8s.io) apiVersion: [rbac.authorization.k8s.io/v1](http://rbac.authorization.k8s.io/v1) kind: ClusterRole metadata: name: everest-pre-upgrade-hook-role rules: - apiGroups: [“”] resources: [“pods”, “services”, “namespaces”] verbs: [“get”, “list”] - apiGroups: [“[apiextensions.k8s.io](http://apiextensions.k8s.io)”] resources: [“customresourcedefinitions”] verbs: [“get”, “list”] - apiGroups: [“apps”] resources: [“deployments”] verbs: [“get”, “list”] - apiGroups: [“”] resources: [“nodes”] verbs: [“get”, “list”] - nonResourceURLs: [“/version”, “/healthz”, “/apis”, “/api”] verbs: [“get”] but stuck with next issue: Checking requirements for upgrade to version 1.8.0 ❌ an empty namespace may not be set when a resource name is provided could not retrieve subscriptio ---
**Author:** ![pstekunov](https://sea1.discourse-cdn.com/flex019/user_avatar/forums.percona.com/pstekunov/32/20328_2.png) [@pstekunov](https://forums.percona.com/u/pstekunov)\ **Post date:** [August 8, 2025, 8:41am UTC](https://forums.percona.com/t/upgrade-to-everest-1-8-0-fails-with-context-canceled-due-to-hanging-pre-upgrade-hook/39026/3 "2025-08-08T08:41:39Z")
> The same with everestctl: > > $ everestctl upgrade --in-cluster --dry-run --version-metadata-url=[https://check.percona.com](https://check.percona.com) > ❌ an empty namespace may not be set when a resource name is provided > could not retrieve subscription ---
**Author:** ![Diogo\_Recharte](https://sea1.discourse-cdn.com/flex019/user_avatar/forums.percona.com/diogo_recharte/32/8004_2.png) [@Diogo\_Recharte](https://forums.percona.com/u/Diogo_Recharte)\ **Post date:** [August 8, 2025, 4:09pm UTC](https://forums.percona.com/t/upgrade-to-everest-1-8-0-fails-with-context-canceled-due-to-hanging-pre-upgrade-hook/39026/4 "2025-08-08T16:09:13Z")
For anyone still interested in this, @pstekunov got his answer on github: [everest 1.8.0 upgrade error · Issue #1568 · percona/everest · GitHub](https://github.com/percona/everest/issues/1568) The workaround is to remove the `app.kubernetes.io/managed-by=everest` label from the `everest-system` namespace: ```auto kubectl label namespace everest-system app.kubernetes.io/managed-by- ```