NetApp ONTAP Upgrades

Summary

This video tutorial covers the essentials of upgrading NetApp's ONTAP system, focusing on the recommended upgrade method, Automatic Non-Disruptive Upgrade (ANDU). The process includes checking platform and software compatibility, using Active IQ's Upgrade Advisor, and following the System Manager wizard for a smooth upgrade. The tutorial also discusses the difference between rolling and batch upgrades, and the need for manual intervention in certain cases like MetroCluster. Additionally, the video highlights potential disruptions during upgrades, especially when using stateful protocols like SMB. It concludes with a guide on firmware and disk qualifications for a successful upgrade.

Takeaways

• 😀 Always check the relevant documentation before performing an ONTAP upgrade, including the upgrade advisor in Active IQ, release notes, and the express guide.
• 😀 Use Upgrade Advisor in Active IQ to generate a personalized upgrade plan, which will guide you through the upgrade process step by step.
• 😀 Verify that your hardware platform is supported for the new ONTAP version and ensure the compatibility of your switches and storage components.
• 😀 Review and ensure that the client operating systems and host utilities are compatible with the new ONTAP version before performing the upgrade.
• 😀 Make sure your current ONTAP version can be directly upgraded to the target version or if intermediate upgrades are necessary (e.g., from 9.2 to 9.3 first).
• 😀 Check that system limits are not exceeded (e.g., number of snapshots, CPU and disk utilization below 50%) to avoid issues during the upgrade process.
• 😀 Download the ONTAP software image and, starting with ONTAP 9.4, you can directly copy the image from a laptop to the storage system without needing an external server.
• 😀 The automatic non-disruptive upgrade (NDU) method is the recommended approach for upgrading ONTAP using a wizard in System Manager, with three stages: Select, Validate, and Upgrade.
• 😀 There are two upgrade types: Rolling Upgrade (one node at a time) and Batch Upgrade (faster, supported on clusters with 8 or more nodes).
• 😀 A rolling upgrade is done one node at a time, with each node’s HA partner taking over the storage, while a batch upgrade upgrades multiple nodes simultaneously, cutting the time in half for large clusters.
• 😀 If you're using a MetroCluster, a manual upgrade must be performed instead of NDU, with more involved pre-upgrade checks performed via the command line.
• 😀 Disruptions during the upgrade can occur if you’re using stateful protocols like SIFS, but stateless protocols like NFSv3, Fibre Channel, and iSCSI are less susceptible to interruptions.
• 😀 Firmware upgrades for system components like disks and shelves are bundled with the ONTAP upgrade, but you must manually update the Disk Qualification Package (DQP) before adding new drives or upgrading firmware.

Q & A

What is the first step to take before performing an ONTAP upgrade?

The first step is to use the Upgrade Advisor in Active IQ to generate an upgrade plan. This plan is available in PDF or Excel format and outlines the upgrade process for your specific system.

What should you check in the release notes before upgrading ONTAP?

You should review the release notes for the version of ONTAP you are upgrading to. This will help you identify any potential issues that could affect your environment.

What are some of the manual checks to perform before starting an ONTAP upgrade?

Before starting an upgrade, you should check the hardware platform for compatibility, verify network switches are supported, confirm that the client operating system and host utilities are compatible with the new ONTAP version, and ensure that the cluster does not exceed system limits like CPU or disk utilization.

Why is it important to check CPU and disk utilization before an upgrade?

It is important to ensure that CPU and disk utilization are below 50% to avoid overloading the HA partner node during the upgrade process. During the upgrade, the storage has to fail over to the HA partner, doubling the workload temporarily.

How does an automatic non-disruptive upgrade (NDU) differ from a manual upgrade?

An NDU is automated through a wizard in System Manager, simplifying the upgrade process, while a manual upgrade requires performing checks manually via the command line and does not offer the same automated validation and support.

What is the purpose of the 'validate' stage in the NDU wizard?

The 'validate' stage in the NDU wizard checks the health of the cluster to ensure that it is ready for the upgrade. If any issues are found, the wizard provides a report with remedial actions that need to be addressed before continuing with the upgrade.

What is the difference between a rolling upgrade and a batch upgrade?

In a rolling upgrade, each node is upgraded one at a time, with the HA partner taking over the storage during the upgrade process. In a batch upgrade, the cluster is divided into two batches, with multiple nodes being upgraded simultaneously, reducing the overall time for the upgrade.

What type of clusters are suitable for a batch upgrade?

Batch upgrades are supported only on clusters with eight or more nodes. This allows for upgrading multiple nodes at once, speeding up the overall upgrade process.

What should you do if you're using the MetroCluster configuration for an upgrade?

If you're using a MetroCluster configuration, you cannot use NDU. Instead, you must perform a manual upgrade with a rolling upgrade method and manually validate the cluster health before starting the process.

How can the use of stateful and stateless protocols affect an upgrade?

Stateless protocols like NFS v3, Fibre Channel, and iSCSI are less affected by disruptions during an upgrade because they don't maintain session states. However, stateful protocols like SIFS may experience disruption, so it's recommended to perform the upgrade during a maintenance window or ensure sessions are ended before proceeding.