(In reply to Thorsten Kukuk from comment #5) > With flannel the error is for me much seldom than with weave. > But it looks like the best way to find out if the cluster is affected or not > is: > run a busybox container and use nslookup to resolve a host. On an affected > cluster you will run into a timeout (temporary failure in name resolution), > else you should get immediately a response. > > A second kubernetes cluster is running fine for me without the issues. > > Between, kured is also broken since the last systemd update is incompatible > ... Again, not a kured issue for me as it affects other services. What is the configuration on your unaffected cluster ? Is it a fresh install ?