Details
-
Bug
-
Resolution: Done
-
Major
-
None
-
4.10.z
-
-
-
Important
-
No
-
SDN Sprint 241, SDN Sprint 242
-
2
-
Rejected
-
False
-
Description
OpenShift 4.10.60 cluster is deployed over Azure with SDN plugin. The egress IP isn't getting attached to the node.
oc project openshift-cloud-network-config-controller oc logs cloud-network-config-controller-5844b4d4b9-vkcqc ... 2023-07-19T13:08:12.903967175Z E0719 13:08:12.903877 1 controller.go:165] error syncing '<masked-node-name>': error retrieving the private IP configuration for node: <masked-node-name>, err: compute.VirtualMachinesClient#Get: Failure sending request: StatusCode=0 -- Original Error: context deadline exceeded, requeuing in node workqueue 2023-07-19T13:08:22.905571458Z E0719 13:08:22.905505 1 controller.go:165] error syncing '<masked-node-name>': error retrieving the private IP configuration for node: <masked-node-name>, err: compute.VirtualMachinesClient#Get: Failure sending request: StatusCode=0 -- Original Error: context deadline exceeded, requeuing in node workqueue...
SDN logs 2023-07-19T12:47:19.686716011Z E0719 12:47:19.686662 1946 egressip.go:260] Ignoring invalid HostSubnet <masked-node-name> (host: "<masked-node-name>", ip: "10.x.x.x", subnet: "10.x.x.0/23"): related node object "<masked-node-name>" has an incomplete annotation "cloud.network.openshift.io/egress-ipconfig", CloudEgressIPConfig: <nil>
As expected the `cloudprivateipconfig` isn't getting created. Even the URL `https://management.azure.com/` is accessible from the pod and nodes. I will be attaching the cluster must-gather along the PCAP file captured from primary interface of all master nodes where we deleted the `cloud-network-config-controller` pod to verify the request going to `https://management.azure.com/` URL.
