Users of the NetApp® ONTAP® storage operating system rely on its powerful SnapMirror® replication technology to replicate volumes between ONTAP systems. With the introduction of the volume migration feature, users can now migrate their ONTAP hosted volumes to Google Cloud NetApp Volumes (NetApp Volumes) by using SnapMirror based replication.
Volume replication
The new feature enables administrators to migrate flex volumes from ONTAP systems like NetApp AFF and FAS, Cloud Volumes ONTAP, and Amazon FSx for NetApp ONTAP to NetApp Volumes. It’s designed to be robust, correct, and high performing, and to minimize required downtime during cutover of users and workloads.
A prerequisite to volume migration is the establishment of a network connection between the source ONTAP system and the Google network (VPC) that the NetApp Volumes are attached to. Typical approaches to achieving this connection are establishing temporary VPN connections or interconnects. This part may require support from network administrators.
Once network connection is available, setting up a volume migration consists of the following phases.
Authentication
In the authentication phase, administrators of the source ONTAP system grant NetApp Volumes permission to fetch volumes from a storage virtual machine (SVM) by doing a cluster and SVM peering.
Baseline transfer
A baseline transfer makes a SnapMirror snapshot of the source system and replicates all of the used data, including all previous snapshots, to the destination volume. Depending on the network speed between source and destination and the amount of data, this process can take several hours, or even days. Meanwhile, the source volume is available and can be used to read and write data.
After the baseline transfer finishes, the destination volume, which contains the data of the SnapMirror snapshot, becomes accessible read-only.
Incremental transfers
While the baseline transfer was in progress, a lot of time may have passed, and a lot of data may have been modified on the source system. A major objective of every migration is to have the data of source and destination in sync and to not lose a single byte.
Volume migration schedules hourly incremental transfers. During an incremental transfer, a new snapshot is taken on the source system, the changes between the new and the previous snapshot are calculated, and only changed data is transferred. Depending on the amount of data changed since the baseline snapshot was made, an incremental transfer will be considerably faster.
With multiple incremental transfers, the amount of data to be transferred is reduced with every iteration, until the change set is minimal. That is the time to schedule downtime for the cutover.
Cutover
During cutover, the administrator moves workloads and/or users access from the ONTAP source volumes to the NetApp Volume destination volumes.
First, all write access to the source system is stopped to avoid data changes. This usually involves shutting down the applications and blocking users. A final incremental transfer copies the latest changes from source to destination. Because the change set has already been minimized by doing incremental transfers, this step finishes quickly.
Next, the migration is stopped. The destination volume becomes read-writable, and all of the workloads and users can be redirected to start using the new NetApp Volumes destination volume. This process involves remounting/remapping shares on the client side.
Cleanup
After the workloads are switched, the migration can be finished, freeing up the migration resource.
Why SnapMirror?
NetApp SnapMirror has many advantages over conventional data copy approaches:
- It works over any IP network and is resilient to network issues. It works over a wide range of network speeds and latencies.
- It only copies used data.
- After an initial (baseline) data transfer, it can do incremental forever transfers of changed data only.
- Calculation of changes for incremental transfers is very fast and independent of the type of data stored in the volume.
- Transfers retain storage efficiency. If the source volume contains deduplicated and/or compressed data, these efficiencies are carried over, reducing the amount of data to transfer.
- Transfers are in-flight encrypted.
- The source volume can be used without noticeable performance impact.
- The destination volume can be used read-only after the baseline transfer finishes.
- It carries over all data, including metadata-like complex access control lists and locked files.
Because of these attributes, SnapMirror is the simplest and most reliable way to transfer volumes between ONTAP systems, even over different geographies.
Feature availability
Volume migration is now available as allow-listed general availability. Contact our Google Cloud specialists to start planning your migration from on-premises ONTAP to NetApp Volumes today.
