Scenario:

• pod <podA> on node <nodeA> in namespace <nsA> accesses a host networked pod <podB> on node <nodeB>
• EgressIP <eipA> is assigned to namespace <nsA>
• Traffic from <podA> to <podB> breaks

Facing timeout errors if the pod is on the same node as the EgressIP and trying to access the Kubernetes API.

~~~~~~~~~~~~~~~~~

EgressIp of the project "manorr-test1": 10.173.128.20

Source node ip where egress IP is attached:  ocp4workerlab101.lmv.lm.se (10.173.0.21)

~~~~~~~~~~~~~~~~~

Below are the links to all the data required for troubleshooting:

Must-gather, sosreport, tcpdump:

https://attachments.access.redhat.com/hydra/rest/cases/03448595/attachments/19856138-beee-4dd1-8ae5-faf92596478a?usePresignedUrl=true

 ovntrace: https://attachments.access.redhat.com/hydra/rest/cases/03448595/attachments/7e9b23ef-2407-4a5f-8726-5e84c071af51?usePresignedUrl=true

OVN NB dump:

https://attachments.access.redhat.com/hydra/rest/cases/03448595/attachments/dd3469bc-6010-46ef-bd6a-a00026883399?usePresignedUrl=true

As shared by the customer, they are facing the same issue addressed in the below Bugzilla which is fixed in 4.11, but right now customer is on the 4.11.32 version,  so need to address the same issue via this bug in 4.11.32

https://bugzilla.redhat.com/show_bug.cgi?id=2070929