Hendrik Ferber
Hendrik Ferber
We run into a similar issue. DB Switchover during node-maintenance (cordon worker nodes). After node-maintenace (uncordon) Standby came back online but constantly crashlooping with ``` ... {"level":"info","ts":"2024-07-26T09:04:17Z","logger":"pg_rewind","msg":"pg_rewind: connected to server","pipe":"stderr","logging_pod":"database-prod-1"}...
@b-thiswatch The filesystem was already full but if I remember correctly the space was occupied by `pg_wal` directory. We end up deleting the database-prod-1 pod and pvc to create a...
i ran this as cluster-admin and got the error. I also tried to allow all ingress connection from the kube-system namespace which also not working. It only works when i...
@sxd Thanks for your support. Don't know which permission i need to add to the cluster-admin user? Currently the status command only works when i allow all ingress traffic towards...
After some deep-dive i can reproduce it in a 2 node minikube cluster. With cnpg plugin version 1.23.3 the status command use a proxy instead of exec into the pods...
As we use calico we implemented the following globalnetworkpolicy to allow the cnpg plugin proxy traffic to every cnpg cluster on the status port ```yaml apiVersion: projectcalico.org/v3 kind: GlobalNetworkPolicy metadata:...
@till can you show your network-policy? Which CNI do you use?
Can you double check that the networkpolicy match the right labels? Is there any default-deny policy in place? Which cnpg plugin version do you use? > Using Calico, and the...
upgrade to v17.0.2 in distributed mode also failed with: ``` failed: cannot patch "minio" with kind StatefulSet: StatefulSet.apps "minio" is invalid: spec: Forbidden: updates to statefulset spec for fields other...