Skip to main content

Considerations for upgrading controller hardware

Contributors netapp-pcarriga

To plan for the upgrade, you must familiarize yourself with the general upgrade considerations. If necessary, contact technical support for recommendations and guidance specific to the configuration of your cluster.

Requirements and limitations

Depending on your environment, you need to consider certain factors before you start your upgrade. Get started by reviewing the table below to see the requirements and limitations you need to consider.

Note Before starting your controller upgrade procedure, you must review all the questions listed in the table below.
Ask yourself…​ If your answer is yes, then do this..

Am I combining different controller platform models in a cluster?

The controllers in a HA pair must be two AFF, FAS, or ASA models.

Do I have different ONTAP versions running on the original and new nodes?

  1. Check the ONTAP versions and patch levels supported by the original and the new nodes.

  2. NetApp recommends, when possible, that you have the same ONTAP version running on the original and new nodes. If this is not possible, upgrade the ONTAP version on the original nodes to the maximum supported version so that the version difference between original and new nodes is NOT greater than four. For example, ONTAP 9.8 and 9.12.1 is supported; however, ONTAP 9.8 and 9.13.1 is not supported.

Do my systems contain internal drives and am I moving volumes?

  1. Verify that the new nodes have enough storage to accommodate storage associated with the original nodes.

  2. When you upgrade by moving volumes, new nodes are joined to the cluster before the original nodes are removed. You must observe the maximum cluster size.

    Note If you are upgrading an eight-node cluster serving block protocols such as FCP, iSCSI, or FCoE, verify that the new nodes are advertising the LUNs correctly. For more information, see SAN storage management.
  3. Upgrade by moving volumes (a nondisruptive procedure).

Am I moving internal storage or converting the system to a drive shelf?

Am I upgrading a HA pair in a cluster with multiple HA pairs?

Move epsilon to the node of a HA pair not undergoing a controller upgrade. For example, if you are upgrading nodeA/nodeB in a cluster with the HA pair configuration nodeA/nodeB and nodeC/nodeD, you must move epsilon to nodeC or nodeD.

Am I running ONTAP 9.6P11, 9.7P8, or later releases?

NetApp recommends that you enable Connectivity, Liveliness, and Availability Monitor (CLAM) takeover to return the cluster into quorum when certain node failures occur. The kernel-service command requires advanced privilege level access. For more information, see the NetApp Knowledge Base article CLAM takeover default configuration changed.

Beginning with ONTAP 9.8, the kcs-enable-takeover parameter is enabled by default.

Note You can upgrade an integrated system by moving data to new storage (moving volumes) or converting the existing integrated system into a shelf and then migrating it to a new system (moving storage). For example, you can upgrade a FAS2650 to a FAS8300 by converting the FAS2650 controller chassis into a DS224C SAS shelf and attaching it to the FAS8300. In either case, the data migration or converted shelf remains in the same switched cluster.

Systems with internal storage

The following systems have internal storage:

Systems with internal drives

FAS2620, FAS2650, FAS2720, and FAS2750

AFF A150, AFF A200, AFF A220, AFF A250, AFF A700s, and AFF A800

AFF C190, AFF C250, and AFF C800

ASA A150, ASA A250, ASA A800, and ASA AFF A220

  • If your system is not listed above, see the NetApp Hardware Universe to check if it has internal drives.

  • If you have a system with internal storage, you can convert the system to a drive shelf and attach it to a new node in the same cluster.

    Note You cannot convert AFF A700s, AFF A800, AFF C800, or ASA A800 systems to a drive shelf.
  • If you have a system with internal storage or a system with volumes or aggregates on internal SATA drives or SSDs, you can upgrade by transferring the internal storage to a drive shelf that is attached to the new node in the same cluster.

    Transferring the internal storage is an optional task in the workflow for upgrading by moving storage.

Situations where you might need additional steps