I just finished a hair-pulling round of testing to try and figure out what was going on with this CentOS 7 Generic Cloud Image that wouldn’t boot after using Ansible on it to prepare it. As it turns out, the issue seems to have nothing to do with what Ansible is doing. What is happening, is that when I use the Ansible VMWare module to power off the VM and convert it to a template, if you “launch VM from template”, the image won’t come up with an IP. And you get a bunch of shared library errors. You can fix this if you, by hand, run “ldconfig” - and then reboot the VM (or manually do an ifup eth0 followed by cloud-init init). I see same thing happening when I use the Ansible module to export the VM to an OVF before saving it as a template. Apparently VMWare is doing something to screw up this VM. I am about to open a ticket with VMWare over this, but if anyone has experienced this and has any suggestions on how to fix it, please let me know. I don’t believe I saw this happening until they upgraded from vSphere 6.7 to 7.x.