Upgrade directly from Rancher v2.12.1 to v2.13.1 fails

Article Number: 000022276

Environment

A SUSE Rancher v2.12.1 instance upgraded directly to v2.13.1

Situation

After upgrading directly from Rancher v2.12.1 to v2.13.1, new Rancher v2.13.1 Pods crashloop, with error messages of the following format:

2026/01/14 15:07:28 [ERROR] Failed to connect to peer wss://10.42.0.90/v3/connect [local ID=10.42.0.110]: websocket: bad handshake
E0114 15:07:32.031893      43 reflector.go:205] "Failed to watch" err="failed to list *v1.Token: object *v1.TokenList does not implement the protobuf marshalling interface and cannot be encoded to a protobuf message" logger="UnhandledError" reflector="/root/.cache/go/modcache/k8s.io/client-go@v0.34.2/tools/cache/reflector.go:290" type="*v1.Token"
2026/01/14 15:07:33 [ERROR] Failed to connect to peer wss://10.42.0.93/v3/connect [local ID=10.42.0.110]: websocket: bad handshake
2026/01/14 15:07:33 [ERROR] Failed to connect to peer wss://10.42.0.90/v3/connect [local ID=10.42.0.110]: websocket: bad handshake
2026/01/14 15:07:38 [ERROR] Failed to connect to peer wss://10.42.0.93/v3/connect [local ID=10.42.0.110]: websocket: bad handshake

Resolution

Upgrading directly from v2.12.1 to v2.13.1 is against the recommended and supported upgrade path for Rancher. Rancher should be upgraded to the latest patch release of the currently running minor version, before upgrading to the latest patch release of the next minor version. At the time of writing, this means upgrading a SUSE Rancher Prime v2.12.1 instance to v2.12.5, before the upgrade to v2.13.1. With this supported upgrade path the issue is not encountered.