White Papers

Additional resources
15 Dell EMC SC Series: Disaster Recovery for Microsoft SQL Server Using VMware Site Recovery Manager | CML1018
4.1.2 Using Replay Manager with vRDMs
The VMware Virtual Machines extension can be used to create application consistent snapshots of SQL
Server data stored on vRDMs. This backup set will back up all virtual disks (.vmdks) and vRDMs used by the
selected virtual machines. If there are other virtual machines using any of the same datastores, it is
recommended to include those virtual machines in the same backup set as well.
4.1.3 Using Replay Manager with pRDMs
To create application consistent snapshots of SQL Server data stored on pRDM, use the SQL Databases
backup extension. This extension is used inside of the guest to take snapshots of the volumes containing
SQL Server databases. Consider the following when using Replay Manager with pRDMs:
• Use the SQL Databases backup extension to take snapshots of the database volumes on pRDMs.
For other types of volumes, use the appropriate Replay Manager backup extension or use an SC
Series snapshot profile.
• With respect to VMware, the SQL Databases backup extension works with pRDMs only. When using
the SQL Databases backup extension, Replay Manager cannot create application consistent
snapshots of databases that are stored on either .vmdks or vRDMs. If pRDMs are used for one
database, it is highly recommended to use pRDMs for all databases, including the system databases.
4.1.4 Using snapshot profiles
Snapshot profiles can also be used to create snapshots directly on SC volumes. However, those snapshots
will not be application consistent. SQL Server databases may not be recoverable from those snapshots. The
odds of having recovery issues are higher in environments where database files for a given database are
placed on multiples volumes, with write-intensive databases being particularly vulnerable. The risk of recovery
problems can be reduced by using a consistency group to take a snapshot of all database volumes at the
same point in time.
4.2 Choosing a replication strategy
Replicating data enables greater efficiency in disaster recovery, yielding major improvements in RPO and
RTO. In other words, replicating virtual machines to a DR site means that during a DR test or actual DR
execution, VMs can simply be powered on at the DR site as opposed to rebuilding or restoring operating
systems and their applications, which is where much of the RTO improvement is introduced.
VMware SRM leverages the replication of storage in its automated recovery workflow. At the time of this
publication , two replication approaches are supported: vSphere host-based replication, which is bundled with
vSphere 5.1+ and replicates at the virtual machine level, or storage vendor array-based replication which
occurs at the datastore or volume level. This paper focuses on SC Series replication and the integrated
solutions that are available with it.
SC Series supports a variety of replication configuration options. Between arrays, replication can be carried
out through FC, iSCSI, or a combination of both.
4.2.1 Asynchronous
Asynchronous replication occurs on a per-volume basis and the replication interval is typically based on the
volume’s snapshot schedule or the frequency at which snapshots are created for a volume. When
asynchronous replication is configured for a volume, each time a snapshot completes it is replicated to the
target system. As an example, if a snapshot is created on a volume every hour, asynchronous replication is