For your existing deployment, if you deployed multiple 20.10.X virtual appliances in a cluster for failover or high availability, you upgrade the nodes one at a time to 21.08.

Expect some downtime during upgrade and plan the timing of your upgrade accordingly. Reindexing Elasticsearch in the first node is a long process that can take over an hour to complete. See Migrating Elasticsearch Data from Workspace ONE Access 19.03.X and 20.01.X before Upgrading to 21.08.


  1. Take snapshots of the database and the service nodes. See KB article 2032907, Managing snapshots in vSphere Web Client.
  2. Remove all nodes except one from the load balancer.
  3. Upgrade the node that is still connected to the load balancer.

    The first node that you update must reindex the Elasticsearch indices. When you see the warning message asking you to run the /usr/local/horizon/update/reindexingIndices.hzn script, enter y. When you upgrade the other nodes in the cluster, enter n to this question. The reindex script is applied only on the first node.

    Follow the process for an online or offline upgrade, as described in Online Upgrading to Workspace ONE Access 21.08 or Upgrading Workspace ONE Access Offline.
    Important: Expect some downtime during the upgrade process.
  4. After the node is upgraded, leave it connected to the load balancer.
    This ensures that the Workspace ONE Access service is available while you upgrade the other nodes.
  5. Upgrade the other nodes one at a time.
  6. After all the nodes are upgraded, add them back to the load balancer.