apologies if this is the correct place to report, but when my dev cluster
(set to the beta channel) upgraded from 1122.1.0 to 1153.3.0 it would no
longer boot. systemd hung on the cloudinit step and no user action could
remedy it. i had to resort to editing grub with 'init=/bin/bash' so i
could get a shell, edit /var/lib/coreos-install/cloud-config file to delete
the coreos update section, reboot into 1153.3.0, rollback to the previous
usr partition, re-ansible the host with stable channel instead of beta, and
reboot again into 1122.1.0. running on vmware using the ova release. all
three of my etcd hosts did the same exact thing so it was repeatable.
