Upgrading KVM Hosts

During the upgrade process, you must migrate all virtual machines to another host in your environment. Doing so minimizes the downtime of virtual machines in your environment. After the upgrade, you can reattach the KVM host to the engine and migrate the virtual machines back to the host.

To upgrade your KVM hosts, you must update them to the latest version of 4.4 before you complete the engine or self-hosted engine upgrade. You also must upgrade any Oracle Linux 7 KVM hosts to Oracle Linux 8. See Before You Begin.

Attention:

When installing or reinstalling the host’s operating system, Oracle strongly recommends that you first detach any existing non-OS storage from the host to avoid potential data loss from accidental initialization of these disks.

Verify the 4.5 engine is installed and running.
(Optional) Verify that all data centers and clusters in the environment are at the same compatibility level.
Pick a host to upgrade and migrate the host’s virtual machines to another host in the same cluster. Any CPU-pinned virtual machines must be shut down and booted into another available KVM host before live migration.
You can use Live Migration to minimize virtual machine downtime. See Migrating Virtual Machines between Hosts.
For ULN registered hosts or if you are using Oracle Linux Manager, ensure you have subscribed to:
- ol8_x86_64_ovirt45
- ol8_x86_64_ovirt45_extras
Install the Oracle Linux Virtualization Manager Release 4.5 package:
```
# dnf install oracle-ovirt-release-45-el8
```
If your host is running UEK R7:
1. Install the Extra kernel modules package.
```
# dnf install kernel-uek-modules-extra
```
2. Reboot the host.
  
  Important:
  If the KVM host you rebooted is in a hyperconverged self-hosted engine environment, review the following information before you continue the upgrade process.
  - After the reboot a KVM host, the glusterd service takes time to check devices and prerequisites. This can take a significant amount of time to complete because it varies based on the number of volumes. And, the volume size influences whether a self-healing process starts.
  - Before attempting to manually activate LVM volumes or services or start healing a volume, check the Gluster logs for errors and failures.
  - Make sure all Gluster storage can be connected: hosted-engine --connect-storage.
  - Check that Gluster volumes and ovirt-ha-agent/ovirt-ha-broker services are in good status on all the KVM nodes:
```
gluster volume status
systemctl status ovirt-ha-agent
systemctl status ovirt-ha-agent
```
Upgrade the KVM host to 4.5:
1. In the Administration portal, go to Compute and then click Hosts.
2. In the Hosts pane, click a blank or non-linked cell for a host to select it.
3. Click Installation and then Check for Upgrade.
4. From the Upgrade Host window, click OK.
  
  The engine checks the KVM host to see if it requires an update.
5. Using your mouse, hover over the icon next to the host name to see if an update is available.
6. To proceed with the update, click Installation and then Upgrade.
7. From the Upgrade Host window, click OK to begin the update process.
Repeat these steps to upgrade the rest of the KVM hosts in the same cluster, one by one, until all are running 4.5.
Update the compatibility version to 4.7 to use all features of release 4.5.

See Changing Data Center and Cluster Compatibility Versions After Upgrading.
Important:
After the engine and KVM hosts are upgraded and verified, remove the release package for version 4.4 on engine and all KVM hosts:
```
dnf remove oracle-ovirt-release-el8
```

Rebooting KVM hosts in a hyperconverged self-hosted engine environment

After the reboot a KVM host, the glusterd service takes time to check devices and prerequisites. This can take a significant amount of time to complete because it varies based on the number of volumes. And, the volume size influences whether a self-healing process starts.