Description of problem: When setting a custom MTU via CNO, the physical NIC is not configured to fit the MTU size. Related to https://bugzilla.redhat.com/show_bug.cgi?id=1846485 Maybe the fix is just adding a quote in the documentation explaining that the dhcp server needs to specify the custom MTU setting as well Version-Release number of selected component (if applicable): 4.5 IPI baremetal but I guess it would be applicable to 4.X in any environment How reproducible: Install an environment with custom MTU without setting any dhcp option related to MTU Steps to Reproduce: 1. 2. 3. Actual results: ``` ... 6: eno4: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc mq master ovs-system state UP group default qlen 1000 ... 8: ovn-k8s-mp0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 8900 qdisc noqueue state UNKNOWN group default qlen 1000 ``` Expected results: ``` ... 6: eno4: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 9000 qdisc mq master ovs-system state UP group default qlen 1000 ... 8: ovn-k8s-mp0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 8900 qdisc noqueue state UNKNOWN group default qlen 1000 ``` Additional info: https://access.redhat.com/solutions/5736191
> Maybe the fix is just adding a quote in the documentation explaining that the dhcp server needs to specify the custom MTU setting as well It's not even "as well". You *must* configure the DHCP server so that the nodes will get the right MTU, and then once you do that, there is no need to configure ovn-kubernetes's MTU because it will use the correct MTU automatically. *** This bug has been marked as a duplicate of bug 1877570 ***