-
Bug
-
Resolution: Duplicate
-
Normal
-
None
-
4.18
-
None
-
False
-
Description of problem:
after ovnkube pods restarted. UDN Secondary L3 traffic cannot be accessed from different nodes
Version-Release number of selected component (if applicable):
build 4.18,openshift/api#1997
How reproducible:
Steps to Reproduce:
1. create UDN with Secondary role
2. create 2 pods on different nodes
$ oc get pods -owide NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES secondary-pod-1 1/1 Running 0 9m53s 10.129.2.12 ip-10-0-18-206.us-east-2.compute.internal <none> <none> secondary-pod-3 1/1 Running 0 10m 10.131.0.18 ip-10-0-60-142.us-east-2.compute.internal <none> <none>
3. check pod2pod connection with secondary-net interface
$ oc exec secondary-pod-3 -- ip a 3: secondary-net@if43: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 9000 qdisc noqueue state UP group default link/ether 0a:58:14:65:03:04 brd ff:ff:ff:ff:ff:ff link-netnsid 0 inet 20.101.3.4/24 brd 20.101.3.255 scope global secondary-net valid_lft forever preferred_lft forever inet6 fe80::858:14ff:fe65:304/64 scope link valid_lft forever preferred_lft forever $ oc exec secondary-pod-1 -- curl --interface secondary-net -s 20.101.3.4:8080 --connect-timeout 5 Hello OpenShift!
4. restart all ovnkube pods and check pods connection
$ oc delete pods --all -n openshift-ovn-kubernetes
$ oc exec secondary-pod-1 -- curl --interface secondary-net -s 20.101.3.4:8080 --connect-timeout 5
command terminated with exit code 28
Actual results:
pod2pod connect failed
Expected results:
pod2pod should connect
