after the last updates, I have a big problem with kube-dns. One of the containers (dnsmasq) fails to run end restarts (crashloopbackoff) all the time (log say: Could not start dnsmasq with initial configuration: fork/exec /usr/sbin/dnsmasq: exec format error). If I try to reinstall from scratch (with all the patches prior of the cluster setup), its the same!
The only solution was a rollback to my old setup (started with the version with 1.7.7 and patches till the change to 1.8.7) via restore from snapshot ... This config works without this problems