AWS Storage Blog
Streamline petabyte-scale data migrations with Cloud Write mode on Amazon FSx for NetApp ONTAP
International Data Corporation (IDC)1 predicts that global data creation and consumption will surge to 175 zettabytes (ZB) by 2025. As a result, organizations are looking for swift, dependable, and scalable cloud migration solutions to lift-and-shift their growing on-premises datasets into the cloud. Whether prompted by an imminent lease renewal, termination of a data center, or as part of the initial stages of a technology transformation, your organization may be embarking on a significant migration project.
To illustrate the challenges of large-scale data migrations, consider the case of a leading automotive manufacturer that needed to quickly migrate over 300 TB of data from their old file systems into AWS. Initially, their migration process was taking nearly two weeks to complete because they needed to continually pause the migration to allow file system tiering to complete or risk filling up provisioned capacity. This was causing significant delays in completing the migration and moving into production. However, after implementing the Amazon FSx for NetApp ONTAP Cloud Write feature, they were able to dramatically reduce their migration time to just two days. This was an 85% reduction in overall migration time, which not only increased migration speed, but also enabled them to realize the benefits of AWS much more quickly, such as increased agility, scalability, and cost savings.
This post demonstrates how you can migrate your data to FSx for ONTAP quickly and cost-optimize storage using its Cloud Write feature, as well as provides step-by-step instructions to enable and disable Cloud Write, along with other important considerations. The intended audience is users who are planning on migrating non-ONTAP source file systems into FSx for ONTAP.
Solution overview
Amazon FSx for ONTAP is a fully managed storage service that allows you to launch and run NetApp ONTAP file systems in the AWS Cloud. It provides the familiar features, performance, capabilities, and Application Programming Interfaces (APIs) of NetApp file systems with the agility, scalability, and simplicity of a fully managed AWS service.
Furthermore, FSx for ONTAP offers high-performance file storage that’s broadly accessible from Linux, Windows, and macOS compute instances through the industry-standard Network File System (NFS), Server Message Block (SMB), and Internet Small Computer System Interface (iSCSI) protocols. It enables you to use ONTAP’s widely adopted data management capabilities, such as snapshots, clones, and replication, with the click of a button. In addition, it provides low-cost storage capacity that’s fully elastic and virtually unlimited in size, and supports compression and deduplication to help you further reduce storage costs.
What is Cloud Write?
As of November 2023, FSx for ONTAP file systems include a powerful new volume-level feature called Cloud Write. Cloud Write enables you to bypass writing to the solid-state drive (SSD) storage tier by writing data directly to the file system’s capacity pool storage tier. This streamlines, accelerates, and reduces costs when migrating data into AWS.
FSx for ONTAP storage tiers
An FSx for ONTAP file system has two storage tiers: SSD storage and capacity pool storage. SSD storage is provisioned, scalable, high-performance storage that’s purpose-built for the active portion of your dataset. Capacity pool storage is a fully elastic storage tier that can scale to petabytes in size and is cost-optimized for infrequently accessed data. A single volume within a file system can reside entirely on the SSD tier, the capacity pool tier, or span both depending on the tiering policy configured for the volume. This AWS guide for managing FSx for ONTAP storage capacity includes recommendations for how to best size SSD and capacity pool storage tiers based on your requirements.
Figure 1: is an illustration of the SSD and capacity pool storage tiers within FSx for ONTAP.
Figure 1: FSx for ONTAP storage tiers
Note that you pay for the SSD storage capacity that you provision when creating or expanding a file system. In the preceding figure, this includes the capacity to store vol1-vol4 and the Free space. On the other hand, capacity pool is elastic, and you only pay for the space used.
FSx for ONTAP tiering policies
FSx for ONTAP volumes can be configured to automatically tier data between the SSD storage tier to the capacity pool storage tier as shown in Figure 2: This tiering process runs continuously in the background, and the time it takes varies depending on the configured tiering policy and the amount of data that needs to be moved. If the data being migrated into the SSD storage tier is being written faster than the rate at which the background tiering process can move it to the capacity pool tier, then the migration halts because FSx for ONTAP prioritizes front-end traffic over the background tiering task. Adding to this challenge, if this SSD space fills to >98%, then capacity pool tiering is halted until capacity is freed on the SSD storage tier.
The FSx for ONTAP guide to configuring data tiering policies can help you navigate these options.
Figure 2: FSx for ONTAP volume data tiering
Choose from four tiering policies to govern tiering:
- Snapshot-only: (Default value) Moves cold snapshots to the capacity pool storage tier. The minimum cooling period is two days, with a valid range of 2 –183 days.
- Auto: Moves cold user data and snapshots to the capacity pool storage tier based on your access patterns. The default tiering minimum cooling period is 31 days and applies to the entire volume, for both the active file system and the Snapshot copies. Valid cooling period values are 2–183 days.
- All: (*required for Cloud Write) Moves all user data blocks in both the active file system and Snapshot copies to the capacity pool tier. The tiering minimum cooling period doesn’t apply, because the data moves to the capacity pool tier as soon as the tiering scan runs, and you can’t modify the setting.
- None: Keeps a volume’s data in the SSD storage tier, preventing it from being moved to the capacity pool tier. The tiering minimum cooling period doesn’t apply, because the data never moves, and you can’t modify the setting.
The challenge of migrating large datasets into FSx for ONTAP
Data migrations are challenging due to the complexities of moving large volumes of data while maintaining integrity and avoiding downtime. Users aiming to cut storage costs may decide to archive or delete data, thus adding to migration challenges. Migration speed is crucial as organizations aim to minimize downtime and maintain operational efficiency.
By default, data that you write to your file system is first written to its SSD storage tier, and then (if you have enabled tiering) ONTAP later moves it to your file system’s elastic capacity pool tier. As a result, your file system’s SSD storage tier, which has a limited amount of storage, can fill up as you move data into it.
Note that metadata is always stored on the SSD storage tier for best performance, even if the file data is tiered or migrated directly to the capacity pool tier.
Key benefits of Cloud Write
These benefits are also illustrated in Figure 3: and Figure 4:
- Ability to write data directly to the capacity pool storage tier. This eliminates the risk of halting migrations by filling the SSD storage tier, or the need to expand the SSD storage tier prematurely during migrations.
- Optimize costs by not having to expand the SSD tier to accommodate larger datasets during migrations.
- Cloud Write can be disabled after migrations, and active data can be automatically tiered back to the SSD storage tier.
Figure 3: FSx for ONTAP data migration without Cloud Write
Figure 4: FSx for ONTAP data migration with Cloud Write
Prerequisites
Consider the following requirements before continuing with this solution.
- You must be a file system or storage virtual machines (SVM) administrator to setup Cloud Write
- You must be at the advanced privilege level on the ONTAP command line interface (CLI).
- The volume must be a read-write type volume.
- The volume must have the ALL tiering policy enabled.
- The data written directly to the capacity pool tier when Cloud Write is enabled benefits from inline deduplication, compression, and compaction. However, you do not get background deduplication or compression because these rely on tasks that periodically run on data that is stored in the SSD storage tier.
- The volume cannot be the source of a SnapMirror relationship while Cloud Write is enabled. If so, then you receive the following error.
- Error: command failed: Cloud write is not supported because the volume is the destination or source endpoint of one or more SnapMirror relationships.
- Note that Cloud Write can be disabled after a migration, and then a SnapMirror relation can be established.
- The volume cannot already have existing backups associated with it before enabling Cloud Write.
- Don’t forget that after disabling Cloud Write, you should enable backups for your volume so that you always have an offline, durable copy.
- If you want to migrate data directly into the capacity pool tier using SnapMirror from other ONTAP on-premises clusters or FSx for ONTAP file systems, then enable the ALL tiering policy on the destination volume.
- Only the NFS protocol is supported for data migrations using Cloud Write. SMB-based migrations still write to the SSD tier even if Cloud Write is enabled.
Walkthrough
Enabling or disabling Cloud Write is a direct task.
To enable Cloud Write:
1. Secure Shell (SSH) into your FSx for ONTAP file system to access the ONTAP CLI. To learn how to do this, see the Managing File Systems with the NetApp ONTAP CLI section of the FSx for ONTAP user guide. The ONTAP CLI command prompt should look like the following with your full file system ID:
FsxId0397d7520dc295a75::>
2. Set the privilege level to advanced and select y to accept the warning:
set -privilege advanced
Warning: These advanced commands are potentially dangerous; use them only when directed to do so by NetApp personnel.
Do you want to continue? {y|n}: y
When you are in advanced mode, your ONTAP CLI command prompt should include the *:
FsxId0397d7520dc295a75::*>3. Modify an existing volume to enable Cloud Write mode:
volume modify -vserver <vserver name> -volume <volume name> -is-cloud-write-enabled <true|false>The following example modifies a volume named vol1 in SVM svm1 to enable Cloud Write:
volume modify -vserver svm1 -volume vol1 -is-cloud-write-enabled true
[Job XXXX] Job succeeded: volume modify succeeded
To validate that Cloud Write was enabled, use the following command:
vol show -volume <volume name> -fields is-cloud-write-enabled
vserver volume is-cloud-write-enabled
------- ------------------- ----------------------
fsx cloudwritetestvol01 true
To disable Cloud Write:
1. Log in to your FSx for ONTAP file system’s ONTAP CLI using an SSH client.
2. Set the privilege level to advanced:
set -privilege advanced3. Modify an existing volume to disable Cloud Write mode:
volume modify -vserver <vserver name> -volume <volume name> -is-cloud-write-enabled <true|false>The following example modifies a volume named vol1 in SVM svm1 to disable Cloud Write:
volume modify -vserver svm1 -volume vol1 -is-cloud-write-enabled falseTo validate that Cloud Write was disabled, use the following command:
vol show -volume <volume name> -fields is-cloud-write-enabled
vserver volume is-cloud-write-enabled
------- ------------------- ----------------------
fsx cloudwritetestvol01 falseNetApp’s ONTAP documentation shows how to setup Cloud Write. At the time that this post was created, Cloud Write can now be enabled on both newly created and existing volumes using the most recent version of FSx for ONTAP 9.14.1P5 or higher.
Cleaning up
To avoid incurring charges from the resources deployed in this walkthrough, remember to delete the FSx for NetApp volumes that you may have created to test this feature.
Conclusion
In this post, we demonstrated how with the latest release of FSx for ONTAP, users can now use Cloud Write, a native ONTAP feature that can simplify and accelerate the process of migrating large datasets into AWS. This included showing how Cloud Write allows you to bypass writing data to the SSD storage tier and instead write directly to the file system’s capacity pool storage tier, therefor streamlining the migration process and reducing the risk of filling up the SSD tier.
At the same time, users can benefit from the option of writing data directly into the fully elastic capacity pool storage tier that can scale to petabytes in size and is cost-optimized for infrequently accessed data. By leveraging the capacity pool tier, organizations can avoid the need to over-provision expensive SSD storage during large-scale data migrations, leading to significant cost savings. Additionally, the capacity pool tier supports inline deduplication and compression, further optimizing storage utilization and costs.
After the migration is complete, users have the flexibility to disable Cloud Write and allow the ONTAP tiering policies to automatically move colder data from the SSD tier to the capacity pool tier, ensuring that the most active data remains on the high-performance SSD storage while less frequently accessed data is tiered to the cost-effective capacity pool.
Thank you for reading this post. Leave any thoughts or feedback in the comment section.
Citations
1How IDC’s Industry CloudPath & SaaSPath Surveys Can Inform Your Cloud/SaaS Strategy