-
Bug
-
Resolution: Done-Errata
-
Major
-
4.14
-
None
-
Critical
-
No
-
SDN Sprint 242
-
1
-
Rejected
-
False
-
This is a clone of issue OCPBUGS-18003. The following is the description of the original issue:
—
Description of problem:
Found auto case OCP-42340 failed in ci job which version is 4.14.0-ec.4 and then reproduced issue in 4.14.0-0.nightly-2023-08-22-221456
Version-Release number of selected component (if applicable):
4.14.0-ec.4 4.14.0-0.nightly-2023-08-22-221456
How reproducible:
Always
Steps to Reproduce:
1. Deploy egressrouter on baremetal with { "kind": "List", "apiVersion": "v1", "metadata": {}, "items": [ { "apiVersion": "network.operator.openshift.io/v1", "kind": "EgressRouter", "metadata": { "name": "egressrouter-42430", "namespace": "e2e-test-networking-egressrouter-l4xgx" }, "spec": { "addresses": [ { "gateway": "192.168.111.1", "ip": "192.168.111.55/24" } ], "mode": "Redirect", "networkInterface": { "macvlan": { "mode": "Bridge" } }, "redirect": { "redirectRules": [ { "destinationIP": "142.250.188.206", "port": 80, "protocol": "TCP" }, { "destinationIP": "142.250.188.206", "port": 8080, "protocol": "TCP", "targetPort": 80 }, { "destinationIP": "142.250.188.206", "port": 8888, "protocol": "TCP", "targetPort": 80 } ] } } } ] } % oc get pods -n e2e-test-networking-egressrouter-l4xgx -o wide NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES egress-router-cni-deployment-c4bff88cf-skv9j 1/1 Running 0 69m 10.131.0.26 worker-0 <none> <none> 2. Create service which point to egressrouter % oc get svc -n e2e-test-networking-egressrouter-l4xgx -o yaml apiVersion: v1 items: - apiVersion: v1 kind: Service metadata: creationTimestamp: "2023-08-23T05:58:30Z" name: ovn-egressrouter-multidst-svc namespace: e2e-test-networking-egressrouter-l4xgx resourceVersion: "50383" uid: 07341ff1-6df3-40a6-b27e-59102d56e9c1 spec: clusterIP: 172.30.10.103 clusterIPs: - 172.30.10.103 internalTrafficPolicy: Cluster ipFamilies: - IPv4 ipFamilyPolicy: SingleStack ports: - name: con1 port: 80 protocol: TCP targetPort: 80 - name: con2 port: 5000 protocol: TCP targetPort: 8080 - name: con3 port: 6000 protocol: TCP targetPort: 8888 selector: app: egress-router-cni sessionAffinity: None type: ClusterIP status: loadBalancer: {} kind: List metadata: resourceVersion: "" 3. create a test pod to access the service or curl the egressrouter IP:port directly oc rsh -n e2e-test-networking-egressrouter-l4xgx hello-pod1 ~ $ curl 172.30.10.103:80 --connect-timeout 5 curl: (28) Connection timeout after 5001 ms ~ $ curl 10.131.0.26:80 --connect-timeout 5 curl: (28) Connection timeout after 5001 ms $ curl 10.131.0.26:8080 --connect-timeout 5 curl: (28) Connection timeout after 5001 ms
Actual results:
connection failed
Expected results:
connection succeed
Additional info:
Note, the issue didn't exist in 4.13. It passed in 4.13 latest nightly build 4.13.0-0.nightly-2023-08-11-101506
08-23 15:26:16.955 passed: (1m3s) 2023-08-23T07:26:07 "[sig-networking] SDN ConnectedOnly-Author:huirwang-High-42340-Egress router redirect mode with multiple destinations."
- clones
-
OCPBUGS-18003 Outgoing traffic throughs EgressRouter is broken
- Closed
- is blocked by
-
OCPBUGS-18003 Outgoing traffic throughs EgressRouter is broken
- Closed
- links to
-
RHSA-2023:5006 OpenShift Container Platform 4.14.z security update