On 9 February 2017 at 10:08, Gianluca Cecchi <gianluca.cecchi@gmail.com> wrote:

On Wed, Feb 8, 2017 at 10:59 AM, Gianluca Cecchi <gianluca.cecchi@gmail.com> wrote:
Hello,
what is considered the best way to shutdown and restart an hypervisor, supposing plain CentOS 7 host?

For example to cover these scenarios:
1) update host from 4.0 to 4.1
2) planned maintenance to the cabinet where the server is located and take the opportunity to update also OS packages

My supposed workflow:

- put host into maintenance
- yum update on host
- shutdown os from inside host (because from power mgmt it is brutal power off / power on)
--> should I get any warning from web admin gui in this case, even if the host was in maintenance mode?
- power mgmt -> start from webadmin gui
(or power on button/virtual button at host side?)

I'd add one more step at the beginning, update the cluster's scheduling policy to InClusterUpgrade. That will automatically migrate your VMs to the most up-to-date host, freeing your other hosts for a rolling upgrade of the cluster.

Would be advisable to put inside power mgmt functionality some logic about os mgmt, so for example, if action is restart, first try to shutdown OS and only in case of failure power off/power on?

Thanks in advance,
Gianluca

I see that one of the new features of 4.1 is Host restart through SSH
Power Mgmt --> SSH Mgmt --> Restart

How does it fit in question above?
Can I choose this way after having run "yum update" on host?

Features are made available only after you update all of the hosts in the cluster & the (Hosted-)Engine to the same version. Then you can update the Compatibility Version of the cluster.

http://www.ovirt.org/documentation/self-hosted/chap-Maintenance_and_Upgrading_Resources/

Doug