I have deployed a private docker registry from here :- https://git.app.uib.no/caleno/helm-charts/tree/84725d0e7921fb75860199fc8f52f2ee58e3d0a4/stable/docker-registry with 2 replica sets. Everything works fine till one of the nodes reboots.
After reboots this command on the rebooted node[root@myhostname3 ~]#curl -u someuser:somepass -k https://<docker_registry_svc_ip>:30500/v2/_cataloggives the delayed response(roughly after 2 minutes). Because of this other pods are going in "imagepullbackoff" state. If i put <pod_ip> instead of <docker_registry_svc_ip> then it is giving instant respone. that means my pod is working fine. Most time taking part is loadbalancing from svc to pod. Can someone please help what is wrong here?Sample output of the curl command.
[root@myhostname3 ~]#curl -u testuser:testpassword -k https://10.42.150.3:5000/v2/_catalog{"repositories":["activemq-postgres","atlantic-network-lldp","backup-restore-operator","busybox","cert-manager-acmesolver-amd64","cert-manager-cainjector-amd64","cert-manager-controller-amd64","cert-manager-webhook-amd64","csi-attacher","csi-baremetal-basemgr","csi-baremetal-controller","csi-baremetal-halmgr","csi-baremetal-loopbackmgr","csi-baremetal-node","csi-baremetal-node-kernel-5.4","csi-baremetal-operator","csi-baremetal-scheduler","csi-baremetal-scheduler-extender","csi-baremetal-scheduler-patcher","csi-node-driver-registrar","csi-provisioner","csi-resizer","fluxcd/helm-controller","fluxcd/source-controller","keycloak","kubernetes/calico/cni","kubernetes/calico/kube-controllers","kubernetes/calico/node","kubernetes/calico/pod2daemon-flexvol","library/decks","library/decks-support-store","library/dell-supportassist-ese","library/ese-callback","library/kahm","library/supportassist-ese-notifier","livenessprobe","local-path-provisioner","metallb/controller","metallb/speaker","minideb","minimal-aaa","nfs-server","nginx","phpfpm","platform-operator","postgres-ha-pgpool","postgres-ha-postgresql","registry","samba4k8s","vault-injector","vault-server"]}Solutions Tried so far
- iptables rules removed (
iptables -F) - disabled firewall
- again rebooted the same node