<button class="button button-with-icon cloud-select-dropdown__button button-primary" aria-haspopup="true">Version 1.16 </button>

- [Latest version](https://docs.cloud.google.com/anthos/clusters/docs/on-prem/latest/how-to/vcenter-upgrade-guide)
- [Version 1.16](https://docs.cloud.google.com/anthos/clusters/docs/on-prem/1.16/how-to/vcenter-upgrade-guide.md.txt)
- [Earlier versions (unsupported)](https://docs.cloud.google.com/anthos/clusters/docs/on-prem/latest/version-history#expandable-1)

<br />

This document describes how to prepare for a vCenter version upgrade and
the downtime during upgrade.

## Disable auto-repair

Before the vCenter upgrade,
[disable `autoRepair`](https://docs.cloud.google.com/anthos/clusters/docs/on-prem/1.16/how-to/node-auto-repair.md.txt#disabling_node_repair_and_health_checking_for_a_user_cluster)
on all user and admin clusters. This prevents machines from being recreated
while vCenter is unavailable.

## Downtime during vCenter upgrade

There shouldn't be any downtime for the admin cluster and user clusters running
in the vCenter during upgrade. The clusters remain up the entire time.
However, there will be downtime for vCenter services like vMotion/DRS and
VM creation.

If your workload directly interacts with vCenter, it might have downtime during
the vCenter upgrade due to the vCenter service unavailabilities.

### Major version upgrade

To upgrade vCenter to the next major version, a new Virtual Center Appliance
will be deployed. And then all vCenter services are shutdown while it copies all
data to the new appliance. Once the data is copied, the new appliance reboots
and takes over the identity (hostname/IP/etc) of the old vCenter.

Major version have been seen to have about 30 mins of downtime for the vCenter
API.

### Minor version upgrade

The minor version upgrade will just update packages inside the OS and
then reboot the appliance, which takes shorter time than major upgrade.

For minor updates, we have seen a downtime around 5 mins for the vCenter API.

## Known Issue

Machines might go to `Unavailable` state after a vCenter reboot for versions
\< 7.0U2. More details see the [known issue](https://docs.cloud.google.com/anthos/clusters/docs/on-prem/1.16/known-issues.md.txt#restarting_or_upgrading_vcenter_for_versions_lower_than_70u2).