host ssh key has been changed after full installation reboot

[George Shuklin]

We've has a planned outage for whole OS installation, and after booting back (+few reboots of hosts and instances during that process) many (may be all) instances changed their ssh keys.

OS: havana@ubuntu
cloud-init:
cloud-init 0.7.2-3~bpo70+1
cloud-initramfs-growroot 0.18.debian5~bpo70+1

cloud-init.log in attachment.

[Joshua Harlow (harlowja)]

Can u try to do curl 'http://10.0.2.2//latest/meta-data/instance-id' from inside a VM? It appears that whats happening is that the following datasources are being searched ['NoCloud', 'ConfigDrive', 'AltCloud', 'OVF', 'MAAS', 'Ec2', 'CloudStack', 'None'] and when you rebooted cloud-init was unable to contact that service, then it switchs to using the None datasource, which believes its instance id is 'iid-datasource-none' which is different from the prior instance-id somehow.

Is there anyway you can upload the contents of /var/lib/cloud (possibly before and after the reboots?), because it should likely have not ran this if it already ran before. Can you also upload your /etc/cloud/cloud.cfg

[Launchpad Janitor (janitor)]

[Expired for cloud-init because there has been no activity for 60 days.]

[James Falcon (falcojr)]

Tracked in Github Issues as https://github.com/canonical/cloud-init/issues/2500