I have airgapped cluster and old version 1.26 and calico 1.16, i do upgrade it to version 1.27 and calico 1.26 because we need do sequence upgrade to latest version anyway it is works then i tried do downgrade as rollback test
i reinstall old rpms and reapply old calico yaml so i have now this problem calico node work put controller and coredns not i try recreate calico yaml i clear all crds service acoount and roles but still not work.
coredns logs say timeout and controller
logs
2025-12-25 12:27:28.331 [INFO][1] main.go 109: Ensuring Calico datastore is initialized
2025-12-25 12:27:38.331 [ERROR][1] client.go 261: Error getting cluster information config ClusterInformation="default" error=Get "https://10.96.0.1:443/apis/crd.projectcalico.org/v1/clusterinformations/default": context deadline exceeded
2025-12-25 12:27:38.331 [FATAL][1] main.go 114: Failed to initialize Calico datastore error=Get "https://10.96.0.1:443/apis/crd.projectcalico.org/v1/clusterinformations/default": context deadline exceeded
Events:
Type Reason Age From Message
Normal Scheduled 23m default-scheduler Successfully assigned kube-system/calico-kube-controllers-6d4549c9d4-pl9jd to pre-prodp-02
Normal Pulled 21m (x4 over 23m) kubelet Container image "reg/kube-controllers:v3.16.10" already present on machine
Normal Created 21m (x4 over 23m) kubelet Created container calico-kube-controllers
Normal Started 21m (x4 over 23m) kubelet Started container calico-kube-controllers
Warning Unhealthy 21m (x10 over 23m) kubelet Readiness probe failed: Failed to read status file status.json: open status.json: no such file or directory
Warning BackOff 2m56s (x85 over 22m) kubelet Back-off restarting failed container calico-kube-controllers in pod calico-kube-controllers-6d4549c9d4-pl9jd_