We aren't setting the MTU correctly on OVN-kubernetes. For example, on an AWS node (which has 9001 uplink mtu), pods only have 1400 MTU: bash-5.0# ip link 3: eth0@if178: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1400 qdisc noqueue state UP mode DEFAULT group default link/ether c6:f1:51:81:00:19 brd ff:ff:ff:ff:ff:ff link-netnsid 0 This is an annoyance on AWS, but is a blocker bug for GCP (which has a less-than-1500 uplink MTU.
This is being fixed as part of https://jira.coreos.com/browse/SDN-456
Also see https://jira.coreos.com/browse/SDN-498
https://github.com/openshift/cluster-network-operator/pull/333
this bug has been fixed in 4.4.0-0.nightly-2020-01-13-010304 ip link 1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN qlen 1000 link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00 3: eth0@if25: <BROADCAST,MULTICAST,UP,LOWER_UP,M-DOWN> mtu 8901 qdisc noqueue state UP link/ether 86:f1:6e:80:02:10 brd ff:ff:ff:ff:ff:ff
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHBA-2020:0581