Introduction
In today's dynamic IT landscape, Application Owners and IT Managers face the challenge of migrating critical workloads between storage clusters without disrupting application availability. Traditional migration methods often result in downtime and potential data loss, impacting business operations and customer experience. However, NetApp's SnapMirror active sync (formerly SnapMirror Business Continuity) offers a solution by enabling live application migration with continuous availability, minimizing disruptions, and maximizing efficiency. In this blog post, we will explore how SnapMirror active sync facilitates seamless application migration and provides a robust business continuity solution.
Note: The process being described should work for any application following iSCSI or FC ALUA priorities. However, the examples in this blog will be addressing ESXi with VMFS datastores backed by iSCSI storage connections.
Understanding the Basics
SnapMirror active sync: A powerful data replication and management solution offered by NetApp. It allows businesses to synchronously replicate data between storage clusters in real-time, ensuring high availability and data protection. By leveraging this technology, organizations can seamlessly migrate applications from one storage cluster to another without impacting end-users.
VMware ESXi: A widely used hypervisor that virtualizes server infrastructure, enabling the creation and management of multiple virtual machines on a single physical server. ESXi provides a reliable and scalable platform for virtualization, with features like high availability and centralized management.
Synchronous Data Replication: A method of data replication where changes made to the primary storage system are synchronously mirrored to a secondary storage system in real-time. This ensures consistent and up-to-date data on both primary and secondary systems.
ALUA: Asymmetric Logical Unit Access (ALUA) is a feature in and Fibre Channel (FC) storage that allows for optimized path selection between initiators and targets, ensuring efficient data transfer and load balancing by designating certain paths as optimized (preferred) and others as non-optimized.
- Active Optimized: Active Optimized is a path state in ALUA where the target storage system actively responds to I/O requests from initiators using the most efficient and optimized path, resulting in higher performance and reduced latency.
- Active Non-Optimized: Active Non-Optimized is a path state in ALUA where the target storage system responds to I/O requests from initiators using a path that may not be the most efficient or optimized, potentially leading to suboptimal performance and increased latency.
Migration Source: Current location of datastore storage.
Migration Destination: Post-migration location of datastore storage.
Environment
For this scenario, let us consider a single ESXi host with a single storage network connection and a NetApp AFF array. The ESXi host utilizes iSCSI to access the backend storage. The existing cluster is the migration source whereas the new cluster will be the migration destination. Figure 1 depicts the current architecture, and Figure 2 illustrates the iSCSI connectivity, where we have two "active" paths, one being optimized, to the storage target.
Figure 1) ESXi hosts with NetApp AFF array.
Figure 2) vSphere client showing two “active” iSCSI paths to the storage.
Verify
Before initiating the migration, it is crucial to ensure the health of your ESXi host that utilizes VMFS backed by iSCSI or FC based storage on NetApp arrays. Verify the connectivity and performance to ensure optimal health with no errors. Taking a proactive approach to monitor your infrastructure, including the viability of iSCSI paths and the operational status of NetApp storage, is essential. Check the network health and storage array conditions before proceeding with the migration.
Setup new storage and SnapMirror active sync relationship
To set up your new NetApp HA pair, refer to the detailed setup instructions provided in the ONTAP Hardware Systems Documentation. Once the new system is in place, ensure you fulfill the prerequisites to establish synchronous data replication by protecting it with SnapMirror active sync. Configuration of the ONTAP Mediator is required, and the LUNs being migrated must be included in a Consistency Group (CG). If a preconfigured CG does not exist, the workflows (system manager, CLI) will assist in creating a new one. Figure 3 illustrates the updated architecture with synchronous data replication, while Figure 4 demonstrates the "healthy" and "In sync" status of the SnapMirror active sync relationship in ONTAP System Manager.
Figure 3) ESXi/NetApp architecture with synchronous data replication.
Figure 4) ONTAP System Manager showing SnapMirror active sync relationship.
Data Migration Process
You are now ready to transition your storage workload from the migration source to the migration destination. Start by adding the iSCSI targets of the migration destination to your ESXi host. This can be accomplished by adding an iSCSI target to the "Dynamic Discovery" list on the "iSCSI Software Adapter" for the ESXi host. Rescan the storage, and you will have four paths to the storage target at this point. However, the active/optimized path is still directed towards the migration source. Figure 5 presents the ESXi/NetApp architecture with four "active" paths, while Figure 6 displays the vSphere client reflecting the four "active" iSCSI paths for the datastore.
Figure 5) ESXi/NetApp architecture with four "active" iSCSI paths.
Figure 6) VMware vSphere client showing four "active" iSCSI paths to the storage.
Now that there are four "active" paths and the SnapMirror active sync relationship is "Healthy" and "In sync," you can proceed with transitioning the storage workload to the migration destination. Perform a planned failover to switch the roles of the NetApp clusters. At this stage, the active/optimized path will migrate to the new cluster. Figure 7 illustrates the ESXi/NetApp architecture with four "active" paths, and Figure 8 displays the vSphere client reflecting the four "active" iSCSI paths for the datastore, with the "active/optimized" path now on the new cluster.
Figure 7) ESXi/NetApp architecture with active/optimized path on the migration destination.
Figure 😎 VMware vSphere client showing the “active/optimized” on the migration destination.
Post-Migration
With the workload running on the migration destination, it is now time to perform post-migration cleanup tasks. These tasks involve removing iSCSI connections to the migration source and disabling synchronous data replication. To remove the iSCSI connections from the ESXi host to the migration source, reconfigure the iSCSI Software Adapter. Within the vSphere client, disable the paths to the original target, remove the old targets from the static discovery, remove the old target from the dynamic discovery, and rescan the storage. After verifying the integrity of your applications, you can proceed to remove the SnapMirror active sync configuration, delete the CGs used for migration, and remove the ONTAP Mediator.
While the application’s storage has been successfully migrated, it is important to note a few key items. SnapMirror active sync focuses on data replication and does not replicate configurations such as snapshot policies and Quality of Service (QoS). These items will need to be recreated manually on the migration destination. Additionally, it is worth mentioning that the original LUNs and data on the migration source remain intact. You may include their cleanup as part of your post-migration process. Figure 9 illustrates the ESXi/NetApp architecture utilizing the migration destination, while Figure 10 displays the vSphere client with paths only to the migration destination.
Figure 9) ESXi hosts using the migration destination.
Figure 10) VMware vSphere client showing paths only to the migration destination.
Conclusion
For Application Owners and IT Managers, SnapMirror active sync provides a reliable solution for non-disruptive application migration to maintain continuous availability. By leveraging this technology, organizations can migrate critical workloads between storage clusters without disrupting end-users. With the combination of NetApp's SnapMirror active sync synchronous data replication and ALUA aware applications, businesses can achieve efficient and uninterrupted application migration, ensuring business continuity and enhancing overall operational efficiency.
Learn More
To learn more about SnapMirror active sync and how it can help your organization visit the following:
