kantver - Fotolia

How are recovery points set in vSphere Replication 6?

When malware strikes or a patch goes bad, a regular backup cycle probably isn't the remedy you need when you want to minimize the loss of data.

It is certainly possible to create and retain multiple virtual machine (VM) backups created with backup-specific tools. But backup cycles tend to be long and infrequent; restoring even the most recent backup may incur serious data loss. Tools such as vSphere Replication 6 can capture and protect complete VMs to any virtual SAN, conventional SAN, NAS or local disk in a fraction of the time needed to perform similar backup tasks.

Replication 6 can further speed replication by capturing only the changes or differences since the last replication process. This allows for point-in-time (PIT) copies in intervals as short as 15 minutes -- or as long as 24 hours which more closely emulates traditional backup cycles.

Frequent replication allows admins to protect VMs with multiple recovery points, and this provides a "rollback" capability that traditional backups cannot offer easily. For example, system configuration changes, application patches, malware attacks and other changes may cause errors or disruptions that might not appear for hours. Multiple recovery points can potentially allow VMs to be restored to earlier "known good" states that existed before any questionable changes occurred. This can aid troubleshooting and support faster problem resolution with minimal data loss.

Using the vSphere Web Client to configure VM replication options, administrators are walked through each step in the setup, including replication type, target site, the replication server running the process, replication options and so on. Replication cycles can range from 15 minutes to 24 hours depending on the importance of the workload. Administrators can choose whether to quiesce a VM before replicating it, though it is best to first quiesce the VM to guarantee the workload is in a steady state before replication. By comparison, a common "snapshot" need not quiesce the VM first, and some rapidly changing or transaction-centric workloads may not yield a stable snapshot if not first quiesced.

Also, consider the retention settings. Frequent replications can demand significant amounts of storage -- even when delta differencing is used to conserve storage and speed replication. VMware's vSphere Replication 6 allows for up to 24 recovery points for each VM; administrators can select the number of instances and number of days -- as long as the total number of instances is less than 25. For example, administrators can configure point-in-time settings to save three recovery points per day for seven days (21 total), five recovery points per day for four days (20 total), 12 recovery points per day for 2 days (24 total) and so on.

Dig Deeper on VMware backup software