-
Bug
-
Resolution: Cannot Reproduce
-
Major
-
None
-
4.12.0
-
None
-
CNF Network Sprint 226
-
1
-
Proposed
-
False
-
Description of problem:
2 vrf test cases constantly failed. Same test cases work on top of ocp 4.11.6
[FAIL] CNF VRF Integration: SRIOV, IPAM: dynamic, Interfaces: 1, Scheme: 2 Pods 2 VRFs ip network overlap [It]
{"Node":"Same Node","IPStack":"ipv4"}
[FAIL] CNF VRF Integration: SRIOV, IPAM: dynamic, Interfaces: 1, Scheme: 2 Pods 2 VRFs ip network overlap [It]
OCP version:
Server Version: 4.12.0-0.nightly-2022-10-04-081353
Kubernetes Version: v1.25.0+3ef6ef3
Version-Release number of selected component (if applicable):
How reproducible:
Apply attached resources but adjust it for your env
My env has secondary sriov nic ens1f0
NodeSelector:
kubernetes.io/hostname: helix09.lab.eng.tlv2.redhat.com
oc create -f namespace.yaml
oc create -f sriov-policy.yaml
oc create -f sriovNetworkBlue.yaml
oc create -f sriovNetworkRed.yaml
oc create -f dhcp-nad.yaml
oc create -f pod-dhcp-server.yaml
oc create -f pod-client.yaml
Client pod can not get DHCP address from dhcp server.
>oc describe pod testpod-client -n cni-tests
Warning FailedCreatePodSandBox 30s kubelet Failed to create pod sandbox: rpc error: code = Unknown desc = failed to create pod network sandbox k8s_testpod-client_cni-tests_6363f2ef-56e4-4baa-8519-b80026d870d6_0(9dbc1c1725e2707044f52481a5d6bb77c1aa677163054bf045c48d8f866c4eff): error adding pod cni-tests_testpod-client to CNI network "multus-cni-network": plugin type="multus" name="multus-cni-network" failed (add): [cni-tests/testpod-client/6363f2ef-56e4-4baa-8519-b80026d870d6:test-vrf-sriov-network-blue]: error adding container to network "test-vrf-sriov-network-blue": failed to set up IPAM plugin type "dhcp" from the device "ens1f0": error calling DHCP.Allocate: no more tries
Normal AddedInterface 30s multus Add eth0 [10.128.3.227/23] from ovn-kubernetes
Actual results:
NAME READY STATUS RESTARTS AGE
testpod-k9lfj 0/1 ContainerCreating 0 49s # Client pod
testpod-pvxvl 1/1 Running 0 53s
pod failed to start
Expected results:
client pod has to get ip address from dhcp server pod as it was in ocp 4.11.
Same tests works on top of the same cluster, same nic (Intel), ocp4.11.6
Additional info:
On the dhcp server pod I can see dhcp request from the client pod
DHCP Server
*18:27:36.979390 IP 0.0.0.0.68 > 255.255.255.255.67: BOOTP/DHCP, Request from 20:04:0f:f1:88:a1, length 348
18:27:36.979830 IP 10.255.255.201.67 > 10.255.255.1.68: BOOTP/DHCP, Reply, length 300
18:27:36.980093 IP 0.0.0.0.68 > 255.255.255.255.67: BOOTP/DHCP, Request from 20:04:0f:f1:88:a1, length 360
18:27:36.980250 IP 10.255.255.201.67 > 10.255.255.1.68: BOOTP/DHCP, Reply, length 300*
Test topology:
Resources