Selecting your update method for non-MetroCluster configurations

Based the requirements of your non-MetroCluster configuration, you can update (upgrade or downgrade) a cluster to a different ONTAP release by performing a nondisruptive upgrade or a disruptive upgrade.

Note: This topic describes options for updating non-MetroCluster configurations. If you are upgrading a MetroCluster configuration, see Selecting your update method for MetroCluster configurations.

Nondisruptive and disruptive updates

Nondisruptive upgrade and downgrade procedures perform the operation while maintaining service to clients.

In a disruptive upgrade or downgrade, storage failover is disabled for each HA pair, and then each node is rebooted one at a time. Disruptive upgrades can be performed more quickly than nondisruptive upgrades, and require fewer steps to complete. However, you should not perform a disruptive upgrade unless you can take the cluster offline for the duration of the upgrade. If you are operating in a SAN environment, you should be prepared to shut down or suspend all SAN clients before performing a disruptive upgrade.

Disruptive upgrade or downgrade is always used for single-node clusters.

Automated and manual updates

Automated nondisruptive upgrades (NDU) are the preferred method of upgrading a cluster. With NDU, ONTAP automatically installs the target ONTAP image on each node, validates the cluster components to ensure that the cluster can be upgraded nondisruptively, and then executes the upgrade in the background.

Nondisruptive manual upgrades involve manual steps to confirm the ONTAP configuration on each node and then use the rolling update method to perform the upgrade or downgrade. In the rolling update method, a node is taken offline and updated while its partner takes over its storage. When the node upgrade is complete, the partner node gives control back to the original owning node and the process is repeated, this time on the partner node. Each additional HA pair is upgraded in sequence until all HA pairs are running the target release

Starting with ONTAP 9.2, automatic updates can also be performed on single-node clusters. However, because single-nodes lack redundancy, updates are disruptive.

Choosing the right procedure

You can use the following diagram to determine which procedure you should use, based on the following criteria:

Below the diagram are links to the procedures.

For this upgrade or downgrade procedure Nondisruptive? Manual? See...
Automated No for single-node clusters; yes for all others. No Upgrading an ONTAP cluster using the automated method
Rolling Yes Yes Upgrading or downgrading a cluster nondisruptively by using the rolling upgrade method
Disruptive No Yes Updating an ONTAP cluster disruptively