"reboot" feature for VMs

[fcami]

A "reboot" capability for any of the IPA nodes used in a test suite would be useful.
Example use cases:

• test if client bits (like sssd or autofs) survive a reboot
• test if a reboot affects replication

That reboot feature should be able to wait for ssh to be operational again and wait for a timeout as well.

[pvoborni]

Just by reading the information available in this bug report, I'd say that it doesn't belong to this project because test should drive the act of rebooting and tests should not be aware of underlying CI.

I know that this ticket is driven by the fact that NFS mounts don't get reattached ( hashicorp/vagrant#1587 ) after reboot. Thus the ticket is valid. But from this perspective, I'd say that job of CI is to get the mounts reattached after a reboot. Waiting for having ssh operational is IMO not something the CI should do - this is a domain of testing framework.

[fcami]

Agreed, yes, though I'm not sure CI can detect a reboot.
An alternative implementation from the testing framework's perspective might be:

• save mount information
• reboot
• on reboot, make sure everything is mounted again
  It is not clear yet how fragile this will be.

[f-trivino]

Please, @netoarmando could you investigate if this is applicable or not?