So far, 2023 has been an exciting year for NetApp®. This year's biggest news has been the introduction of our new NetApp AFF C-Series of value all-flash controllers that offer access to flash performance with the economics of spinning media. We released ONTAP® 9.12.1 software for production environments in January and now we are releasing our second major update for on-prem clusters, ONTAP 9.13.1. With this release, exciting new features continue to increase ONTAP’s value.
As part of every ONTAP release, we continually work to improve SnapMirror®, our premier NetApp data protection suite for multi-protocol disaster recovery and backup/archive use cases. With 9.13.1 now available to you, I want to once again take a few words to talk about the key new features and improvements for SnapMirror.
Support for NetApp AFF C-series controllers
First off, let's get the easy one out of the way. As I mentioned, we recently released our new AFF C-Series family of capacity all-flash controllers, and starting with ONTAP 9.12.1 patch1 (ONTAP 9.12.1P1), SnapMirror asynchronous, SVM-DR, and SnapMirror Synchronous, are qualified for deployment on C-Series controllers. SnapMirror Business Continuity as well as our SVM data mobility tool (SVM Migrate) are qualified on C-Series controllers starting with ONTAP 9.13.1.
Support for larger volumes, files, and LUNs
When we released ONTAP 9.12.1 in January, we added the ability to create larger volumes, files, and LUNs. This functionality had already existed on our AFF All-SAN Array (ASA) offerings, but not for our FAS or AFF mixed-protocol controllers. Released as a Public Preview, this new feature was qualified for customer production deployments starting with ONTAP 9.12.1P2. Though not part of our 9.13.1 launch, I wanted to make everyone aware that volume-scoped SnapMirror - both synchronous and asynchronous - will replicate data between FlexVol® volumes up to 300TB and larger file/LUN sizes up to 128TB each.
SnapMirror Asynchronous and SVM-DR
SnapMirror replication of consistency groups
For some time now, ONTAP has supported consistency groups (CGs) to group volumes and LUNs together to take an application-consistent snapshot of multiple volumes and LUNs. ONTAP offers a tiered CG architecture that allows a CG to consist of multiple child CGs. This can be useful for database applications that have multiple volumes for logs and multiple volumes for data. Using this tiered architecture, backup software providers can have different backup plans for each child CG while enabling the backup of an entire application’s data in one step by using the parent CG.
Starting with ONTAP 9.13.1, SnapMirror can use consistency groups to control asynchronous replication to a DR cluster. For this release, only single-tier, or flat, consistency groups are supported for SnapMirror asynchronous replication. Attempting to create SnapMirror relationships that target CGs containing other CGs (parent-child CGs) will result in an error. Other than this parent-child CG restriction, SnapMirror relationships targeting CG sources will support the same functionality currently offered for asynchronous FlexGroup® volume SnapMirror relationships (see Figure 1 below for a comparison of ONTAP tiered CGs and flat CGs).
Figure 1: Differences in CGs used in ONTAP and SnapMirror
ONTAP CLI, REST APIs, or System Manager can be used to create SnapMirror relationships targeting CGs. Here are some examples of creating SnapMirror CG-scoped relationships using CLI and REST:
Dest::> snapmirror create -source-path SVM1:/cg/db1 -destination-path SVM1-Dest:/cg/db1-dest
-cg-item-mappings db-vol1:@db-vol1_dest, db-vol2:@db-vol2_dest, db-vol3:@db-vol3_dest
-policy MirroAndVault -schedule hourly
POST /api/snapmirror/relationships/
'{"source": { "path": “SVM1:/cg/db1", "consistency_group_volumes": “db-vol1, db-vol2, db-vol3"},
"destination": { "cluster.name": "dest_cluster", "path": “SVM1-Dest:/cg/db1-dest", "consistency_group_volumes": “db-vol1_dest, db-vol2_dest, db-vol3_dest"},
"policy": “MirrorAllSnapshots",
"create_destination": { "enabled": "true", "storage_service": { "enabled": "true", "name": "extreme", "enforce_performance": "true" } }
CG support is limited to asynchronous SnapMirror replication of FlexVol volumes.
Improved recovery speed from DR events and DR rehearsals
Several new features contribute to improving the user experience when performing recovery operations...either as part of a DR rehearsal or after an actual DR event. These improvements include:
|
Feature |
Description |
|
Restore process validation |
While this is something that is not visible to the user, SnapMirror now validates each step of the restore process before starting the restore. If it finds any issues, it will report those much quicker and not require the user to wait through a lengthy restore only to get an error in a later stage of the restore process. |
|
Improved DR rehearsal recovery |
SnapMirror will now always try to perform a quick resync when any resync operation runs unless told otherwise.
A standard resync requires a complete rebuild of the data warehouse on the DR site volume before actual data can be transferred. This process retains as much of the storage efficiencies in place on the DR volume as possible. A quick resync prioritizes recovery time over storage efficiency by not performing the data warehouse rebuild. Any lost storage efficiencies can be recovered after the resync of changed data and the primary site is brought back online. |
|
Quick resync support for FlexGroup volumes |
Before this ONTAP release, quick resync was only allowed for FlexVol volumes. ONTAP 9.13.1 and later can now perform quick resync on FlexGroup volumes. Standard resyncs on FlexGroup volumes - which can be massive compared to FlexVol volumes - could take extremely long times to recover...not something you want if you were only performing a DR rehearsal in a limited span time window. |
By the way, if you do perform a standard resync operation, where the data warehouse is rebuilt before data is transferred back to the primary cluster, you can track the resync operation with the data warehouse resync rebuild progress indicator. This indicator was first introduced in ONTAP 9.12.1, but now supports FlexVol volumes, FlexGroup volumes, and SVM-DR resync operations.
SVM-DR supports FlexGroup volumes
Yes! This has been one of our big customer requests: "Why can't I use SVM-DR to protect SVMs that contain FlexGroup volumes?" Well, now you can. Assuming both the source and destination clusters are running ONTAP 9.13.1, SVMs hosting FlexGroup volumes can now be protected by SnapMirror SVM-DR.
SnapMirror Synchronous
As mentioned earlier, SnapMirror is qualified for deployment on the new NetApp AFF C-Series controllers starting with ONTAP 9.12.1P2. On top of that, we have two items that will help customers that are using SnapMirror Synchronous to talk about.
ONTAP version interoperability improvements
This is something that we have been working on for a few months now, and while it didn’t get completed in time for the ONTAP 9.12.1 release in January, we have now completed our testing. Before ONTAP 9.12.1, the source and destination clusters participating in a SnapMirror synchronous relationship had to be within "two major ONTAP versions" of each other. Starting with ONTAP 9.12.1, SnapMirror Synchronous replication is supported between two clusters if the ONTAP version of the source and destination are within 3 YEARS of each other! If the standard 3-year warranty is in effect for both endpoints, then SnapMirror Synchronous replication is supported. This also matches our ONTAP interoperability policy that is in place for asynchronous volume-scoped SnapMirror.
Support for restoring data to a SnapMirror source volume using NetApp NDMP
Starting with ONTAP 9.13.1, NDMP restore operations can be performed against a source volume. This includes support for all operations for both ndmpcopy and dump/restore commands targeting volumes, directories, files, or LUNs (logical unit numbers). The full list is shown here:
|
NDMP Command |
Supported operations |
|
NDMPcopy |
Full Vol restore Sub-dir File-level/LUN Restore to online LUN Incremental Files with streams and ACLs Exclude-list Environment variables. |
|
dump |
Incremental DAR (a bunch of dirs or files) Sub-dir File-level(w/ rename) Files with streams and ACLs Extract restore Audit of backup (restore with NoWrite) Restore w/ list |
Both the source and destination clusters must be running ONTAP 9.13.1.
SVM data mobility (SMV Migrate)
The big news for SVM Migrate is scale, scale, scale!
SVM Migrate can now migrate an SVM hosting up to 200 FlexVol volumes, a 100% increase from previous ONTAP releases, and we can migrate SVMs between larger clusters...clusters of up to 12 controllers (6 HA pairs) ... again a 100% increase in scalability.
Closing thoughts
Yes, 2023 has been an exciting year for NetApp...and we look forward to even more fun things coming for the remainder of the year. I am excited about the new features and capabilities we continue to drive into our core on-premises products, but we are also making extremely cool stuff for hybrid and private/public cloud ONTAP users, and I encourage you to explore these great products on netapp.com.
Download updated Tech Reports for SnapMirror (TR-4015) and SnapMirror Synchronous (TR-4733) for all the latest configuration and best practices recommendations.
As always, let us know what you think about our new SnapMirror features in the comments below. Also, if you would like to nominate ideas for new or improved SnapMirror features, drop us a post in the community forums, Discord, or talk to your NetApp representative.