Planned maintenance and disaster recovery (DR) tests work best when storage and compute move together. If a workload moves to another zone but storage continues serving from the original zone, every I/O request crosses zones. A regional storage pool also needs a tested recovery plan. Without regular testing, teams may discover gaps during a real incident; clients connect to the wrong zone, recovery procedures have not been rehearsed, or compute is not ready in the standby location.
Google Cloud NetApp Volumes now supports zone switch for Flex Unified regional storage pools. With a single action, you can move the pool’s active zone to its replica zone and switch it back later. There is no need to rebuild storage, create another copy of the data, or change how applications access volumes.
This blog explains how zone switch works, what clients experience during a switch, and how to use the Google Cloud console and gcloud CLI. It also walks through two examples: planned maintenance on a Windows Server Failover Cluster and a DR test.
Regional storage pools: active and replica zones
A regional Flex Unified storage pool maintains two synchronous copies of your data in different Google Cloud zones. When you create the pool, you choose an active zone and a replica zone:
- Active zone: serves all volumes in the pool.
- Replica zone: holds the synchronous standby copy, ready in the background.
If the active zone fails, NetApp Volumes automatically moves NFS, SMB, or iSCSI service to the replica zone. Access resumes after a brief I/O pause, without data loss.
How zone switch works
Zone switch lets you initiate this change on demand. When you run a switch:
- Volume access moves from the current active zone to the replica zone.
- The replica zone becomes the new active zone.
- The previous active zone becomes the new replica zone.
Zone switch changes which zone serves the pool; it does not create a new copy of the data. All volumes in the pool switch together, rather than individually. The volumes, protocols, snapshots, and replication relationships remain unchanged.
Expect a brief I/O pause during the switch. The duration depends on your clients and how they handle the path change, so choose a window that can tolerate an interruption. Once the switch completes, file and block services resume from the new active zone.
What your clients experience
Clients do not need to be reconfigured or remounted. The way access moves between zones depends on the protocol:
- SAN clients (iSCSI): Volumes are presented through an active logical interface (LIF) and a standby LIF. During a zone switch, I/O moves to the standby LIF and continues on that path until you switch the pool back.
- NAS clients (NFS and SMB): Volumes are served through a single LIF. During a zone switch, the LIF moves to the other zone, and clients continue using the same address.
- SMB locks: To preserve locks during the switch, configure SMB volumes as continuously available (CA) shares.
Applications retain their existing paths, mounts, and connection settings.
When to use zone switch
- Planned maintenance: Serve storage from the replica zone while you patch, upgrade, or shut down compute in the original active zone.
- DR readiness: Switch to the replica zone, check that applications reconnect and data is intact, then switch back.
- Failover rehearsal: Practise the console or gcloud steps in your recovery plan, and validate storage, networking, and application access together before an incident.
- Compute alignment: Align storage with compute after a planned VM move.
Keep compute close to storage. For the lowest latency, run workload VMs in the pool’s active zone. If clients remain in the original zone after a switch, I/O crosses zones and may increase latency. Include moving or starting compute in the new active zone in your plan.
Before you switch
After a manual zone switch, the pool remains in a switched state until you switch it back to the original zone. During this time:
- Pool upgrades are paused. Storage pool upgrades wait until you switch back.
- Scaling is unavailable. You cannot increase the pool’s capacity, throughput, or IOPS until you switch back.
Before a test or maintenance window, work through this checklist:
- Scale first. Complete any planned capacity, throughput, or IOPS increases. For a longer window, provision headroom for expected growth.
- Choose a suitable window. Allow for a brief I/O pause.
- Prepare SMB volumes. Configure them as continuously available (CA) shares.
- Plan compute placement. Decide when to move or start compute in the target active zone.
- Plan the switch-back. Keep the window short and include switching back in the same plan. Document both directions in your DR procedures.
Note: Zone switch applies to regional Flex Unified storage pools. Before relying on it for another pool configuration, confirm support in the storage pool documentation.
How to switch the active and replica zones
- Go to NetApp Volumes in the Google Cloud console.
- Click Storage pools.
- Open the regional Flex Unified storage pool you want to switch.
- Click Switch active/replica zones.
- Review the current active and replica zones, and the new roles after the switch.
- Click Switch.
- In the confirmation dialog, enter the pool name, then click Switch again.
Once the operation completes, the active and replica zones have swapped roles.
Switching back
When your test or maintenance is complete, use the same console button or run the same switch command again. The zones swap roles once more, returning the pool to its original active zone.
Example 1: Planned maintenance on a Windows Failover Cluster
Windows Server Failover Cluster (WSFC) deployments, such as SQL Server Failover Cluster Instances (FCIs), commonly use nodes in two zones and a regional Flex Unified pool for shared storage. In this example, one node owns the application roles at a time. For the lowest latency, that node should run in the pool’s active zone.
Consider the following two-node cluster:
- Node A runs in zone
us-east4-a, the pool's active zone. It currently owns the cluster roles. - Node B runs in zone
us-east4-b, the pool's replica zone. It is the passive node.
Suppose you need to patch Node A. Failing the FCI over to Node B moves the workload to zone b, but storage continues serving from zone a. I/O then crosses zones throughout the maintenance window. Use zone switch to align storage with the new compute location:
Step | Cluster (compute) | NetApp Volumes (storage) | Storage active zone after step |
|---|
1. Prepare | Confirm the cluster is healthy. | Confirm the pool is healthy and finish any planned scaling. | a |
2. Move storage | | Run zone switch. | b |
3. Move compute | Drain Node A and fail the FCI over to Node B. | | b |
4. Maintain | Patch, upgrade, or restart Node A. | | b |
5. Rejoin | Bring Node A back into the cluster and confirm it is healthy. | | b |
6. Move compute back | Fail the FCI back to Node A. | | b |
7. Move storage back | | Run zone switch again. Upgrades and scaling are available again. | a |
Run steps 2 and 3 back to back, and do the same for steps 6 and 7. This keeps the period when compute and storage are in different zones short. The storage switches and FCI failovers can interrupt I/O, so rehearse the sequence in a non-production cluster to understand how your application responds.
Example 2: A DR test
- Confirm that the pool is healthy and record the current active and replica zones.
- Complete any planned pool scaling and schedule a window that can tolerate a brief I/O pause.
- Start or move test compute in the replica zone.
- Run zone switch and wait for the operation to complete.
- Verify that I/O resumes without remounting or reconfiguring clients: SAN clients use the standby LIF, while NAS clients retain the same LIF address. Check that data is intact and applications behave as expected.
- Record the results and any gaps in your DR plan.
- Switch the pool back when testing is complete. If you deliberately keep the new active zone in place, remember that scaling and upgrades remain unavailable until you switch back.
Why this matters
Regional Flex Unified pools provide synchronous data protection across two zones. Zone switch lets you put that protection to the test.
You can rehearse DR rather than assume failover will work. You can also patch compute in the original zone while keeping storage aligned with the workload. Throughout the switch, you retain the same volumes, protocols, and data services, including NFS, SMB, iSCSI, snapshots, and replication.
For customers running VMware Engine, databases, file shares, or mixed workloads on Flex Unified regional pools, zone switch turns regional availability from a design goal into a capability they can practise and validate.
Get started
- Create or use a Flex Unified regional storage pool with an active zone and a replica zone.
- Run application compute in the pool’s active zone.
- Schedule and run a zone switch in a test or maintenance window.
- Confirm client and application access, then switch back.
To learn more:
- Review Manage your storage pool
- Talk to your NetApp or Google Cloud account team about DR design and zone alignment for your workloads.
Put zone switch into practice during your next maintenance window or DR exercise, and use the results to strengthen your recovery plan.