Closed joelmccoy closed 1 week ago
Additional context: this happened in two different environments (our staging and production env).
The quick fix was to deploy the upgrade. Scale the Keycloak StatefulSet down to 0 then scale the StatefulSet back up to your desired amount
Environment
Device and OS: AWS rke2 1.29 App version: Keycloak 25 -> 26 Kubernetes distro being used: rke2 1.29 Other: uds-core
v0.30
->v0.31
Steps to reproduce
v0.30
->v0.31
with an HA Keycloak DeploymentExpected result
Graceful upgrade of keycloak and uds-core
Actual Result
1 of the 3 keycloak pods stuck in crashloop and kicking the pods did not resolve the issue
Visual Proof (screenshots, videos, text, etc)
Some snippets of noteable pod log errors:
Severity/Priority
Probably high? Essentially anyone who rolls keycloak HA is going to run into this issue on this upgrade it seems
Additional Context
Apparently Keycloak doesn't really support rolling major upgrades according to this comment so we may need to find a workaround to support these upgrades.