-
Story
-
Resolution: Unresolved
-
Major
-
None
-
None
Description of problem:
OVN-kube node crashes with the following errors: F0809 19:25:35.697554 2129731 ovnkube.go:136] failed to start node network manager: failed to start default node network controller: failed to repair Egress Services entries: [running [/usr/sbin/iptables -t nat -D OVN-KUBE-EGRESS-SVC -A OVN-KUBE-EGRESS-SVC -m mark --mark 0x3f0 -m comment --comment DoNotSNAT -j RETURN --wait]: exit status 2: iptables v1.8.8 (nf_tables): Cannot use -A with -D Try `iptables -h' or 'iptables --help' for more information. , running [/usr/sbin/iptables -t nat -D OVN-KUBE-EGRESS-SVC -A OVN-KUBE-EGRESS-SVC -m mark --mark 0x3f0 -m comment --comment "Do not SNAT to SVC VIP" -j RETURN --wait]: exit status 2: iptables v1.8.8 (nf_tables): Cannot use -A with -D Try `iptables -h' or 'iptables --help' for more information. ] I0809 19:25:35.697583 2129731 reflector.go:293] Stopping reflector *v1.Namespace (0s) from k8s.io/client-go/informers/factory.go:150 With https://github.com/openshift/cluster-network-operator/pull/1874 tested.
Version-Release number of selected component (if applicable):
4.14 (pre feature freeze)
How reproducible:
Always
Steps to Reproduce:
1. Install DPU 2 Cluster 2. Upgrade OVN-K and CNO (https://github.com/openshift/cluster-network-operator/pull/1874) to the latest downstream 3.
Actual results:
We see these rules on the host: sudo iptables -t nat -S OVN-KUBE-EGRESS-SVC -N OVN-KUBE-EGRESS-SVC -A OVN-KUBE-EGRESS-SVC -m mark --mark 0x3f0 -m comment --comment DoNotSNAT -j RETURN -A OVN-KUBE-EGRESS-SVC -m mark --mark 0x3f0 -m comment --comment "Do not SNAT to SVC VIP" -j RETURN Thus OVN-kube node crashes.
Expected results:
We should see these rules on the host instead. sudo iptables -t nat -S OVN-KUBE-EGRESS-SVC -N OVN-KUBE-EGRESS-SVC -A OVN-KUBE-EGRESS-SVC -m mark --mark 0x3f0 -m comment --comment DoNotSNAT -j RETURN
Additional info:
https://github.com/ovn-org/ovn-kubernetes/pull/3064/files#diff-bb96390aa4d292c5ff6e2bc554446182b8abd55e12ba2a0290853b92aa4e1155R589 Slack thread: https://redhat-internal.slack.com/archives/CDCP2LA9L/p1691610257057969
- blocks
-
NHE-109 Bluefield-2 in DPU mode: rerun flow tests
- Testing