---
title: Recovering from a failed upgrade
description: You can recover your PingAccess cluster by switching back to the source version if the upgrade fails.
component: pingaccess
version: 9.0
page_id: pingaccess:pingaccess_zero_downtime_upgrade:pa_recovering_from_a_failed_upgrade
canonical_url: https://docs.pingidentity.com/pingaccess/9.0/pingaccess_zero_downtime_upgrade/pa_recovering_from_a_failed_upgrade.html
revdate: February 6, 2023
section_ids:
  about-this-task: About this task
  steps: Steps
---

# Recovering from a failed upgrade

You can recover your PingAccess cluster by switching back to the source version if the upgrade fails.

## About this task

The zero downtime upgrade process creates a set of new folders for the upgraded installation. The pre-upgrade source installation is not affected.

To recover your PingAccess cluster in the event of a failure, you would resume the former installation using these steps.

## Steps

1. Stop any upgraded PingAccess instances.

2. Start the original PingAccess instance on the admin node.

3. Import the engine definitions back into the original PingAccess instance.

4. Start the original PingAccess instances on the engine nodes.

5. Ensure all engines are added to the load balancer configuration.