-
Bug
-
Resolution: Done-Errata
-
Critical
-
None
-
4.14.0
-
No
-
Sprint 240
-
1
-
Approved
-
False
-
Description of problem:
Configure an ingress-controller or a route with a http request header, whose name's length is larger than 255, the cluster will send back "http 500 Internal Server Error" to a client http request
Version-Release number of selected component (if applicable):
4.14.0-0.test-2023-08-07-002636-ci-ln-w7n69q2-latest
How reproducible:
100%
Steps to Reproduce:
1. % oc get clusterversion NAME VERSION AVAILABLE PROGRESSING SINCE STATUS version 4.14.0-0.test-2023-08-07-002636-ci-ln-w7n69q2-latest True False 8h Cluster version is 4.14.0-0.test-2023-08-07-002636-ci-ln-w7n69q2-latest 2. configure route selector for the default ingress-controller, then create an ingress-controller 3. create a pod, svc and expose the route with specified hostname % oc get route unsec-server3 -oyaml apiVersion: route.openshift.io/v1 kind: Route metadata: creationTimestamp: "2023-08-07T02:54:30Z" labels: name: service-unsecure name: unsec-server3 namespace: default resourceVersion: "59125" uid: 08a6e8ed-c1dc-4509-bba2-56a9236cc858 spec: host: unsec-server3.int1.shudi-414g07.qe.gcp.devcluster.openshift.com port: targetPort: http to: kind: Service name: unsec-server3 weight: 100 wildcardPolicy: None status: ingress: - conditions: - lastTransitionTime: "2023-08-07T02:54:30Z" status: "True" type: Admitted host: unsec-server3.int1.shudi-414g07.qe.gcp.devcluster.openshift.com routerCanonicalHostname: router-int1.int1.shudi-414g07.qe.gcp.devcluster.openshift.com routerName: int1 wildcardPolicy: None 4. Configure the custom ingress controller with a request http header whose name length is 256 % oc -n openshift-ingress-operator get ingresscontroller int1 -oyaml | grep -A7 httpHeaders: httpHeaders: actions: request: - action: set: value: header name length test type: Set name: aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa 5. rsh to pod and try to curl the route % oc rsh centos-pod2 sh-4.4# curl http://unsec-server3.int1.shudi-414g07.qe.gcp.devcluster.openshift.com <html><body><h1>500 Internal Server Error</h1> An internal server error occurred. </body></html> 6. rsh to server pod and check the captured packet, the server doesn't receive a packet on port 8080 % oc rsh server3 sh-4.4# tcpdump -i eth0 port 8080 -s 0 -n -v dropped privs to tcpdump tcpdump: listening on eth0, link-type EN10MB (Ethernet), capture size 262144 bytes ^C 0 packets captured 0 packets received by filter 0 packets dropped by kernel sh-4.4#
Actual results:
http 500 error occurs
Expected results:
the http request with the http header inserted by the controller is forwarded to the server
Additional info:
