Can you get the openshift-sdn pod logs?
actually, it looks like the cluster failed before ever even starting openshift-sdn on the nodes...
Just to summarize findings thus far, Russell Teague has bisected the failure down to the delta between 3.11.273 and 3.11.274. There are no code changes between those two builds, however there is a golang build toolchain change[1][2] from go-toolset-1.10.8 to golang-1.13.4. Just for additional context, the delta[3] between the last (3.11.272..3.11.284) errata and this errata shows code changes[4] meant to enable 1.13 so this change seems intentional rather than by mistake. Since restarting the kubelet resulted in the mirror pods being created this seems likely that any fixes will happen in the kubelet. So moving to Node component. 1 - http://download.eng.bos.redhat.com/brewroot/vol/rhel-7/packages/atomic-openshift/3.11.273/1.git.0.7cb6887.el7/data/logs/x86_64/root.log 2 - http://download.eng.bos.redhat.com/brewroot/vol/rhel-7/packages/atomic-openshift/3.11.274/1.git.0.724cbde.el7/data/logs/x86_64/root.log 3 - https://github.com/openshift/ose/compare/v3.11.272-1...enterprise-3.11 4 - https://github.com/openshift/origin/pull/25416
Toolchain was reverted. Closing.