This is an article about Zerto, the award-winning IT Resilience platform I was first exposed to in 2011 and loved so much that I joined the team in 2012. For help with any Zerto-specific terminology, see Zerto’s official product documentation.
Whenever you perform a Live Failover of any Virtual Protection Group (VPG) in Zerto you’re given the option to enable a feature called Reverse Protection. If enabled, the Reverse Protection feature automatically transposes the VPG configuration after recovery is complete so you don’t have to do it manually. The end result is a VPG configured to put things back where they came from.
One step in the Reverse Protection process that surprises some Zerto users is a Delta Sync. They ask what is Delta Syncing, why is a Delta Sync happening when Zerto should have the same data at each site, and more.
For more about the Zerto Delta Sync operation, see this post on the MyZerto forum.
Zerto performs a Delta Sync as part of Reverse Protection because Zerto inherently distrusts that VPG’s source-site data after a Live Failover event. We as people sometimes fall into the trap of assuming the source data must match the recovered data. After all, the application was shut down as part of the failover process, or the failover only took a minute or two! That assumption is incorrect for several reasons.
First, you are not failing over to the current application state because Zerto replicates asynchronously. The data may only be a few seconds old, but a few seconds is not zero seconds. Next, although you may elect to shut down the source application as part of the Live Failover, Zerto does not replicate data created by that application after you initiate the failover process. This is because, from a product design perspective, the Zerto software doesn’t fall into the “best case scenario” trap I mentioned earlier. Between the shutdown and the recovery, any one of dozens of events may have happened to alter the source application data, storage configuration, or the storage array itself. Zerto assumes a worst-case scenario, that the source and recovery site data may not be the same, and then validates this assumption using a Delta Sync.
If you do want a perfectly-synced application recovery, you want to use the Move operation instead of Live Failover. As with a Live Failover, the software assumes the worst and performs a Delta Sync (assuming you don’t disable Reverse Protection as part of the Move). Unlike a Live Failover, a Move will synchronize any data created by that application in the last moments before it is completely shut down and then “moved” to the recovery site.
Sean M Masters 