"seeding" replication jobs

[tim.vogel@xtium.com]

We have some established replication jobs going from a remote site to our data center over a leased line. The "old" process does not use the Veeam product for replication. We now have Veeam licenses and would like to "convert" to Veeam. What is the best way to "seed" a new Veeam replication job if we already have copies of the VM on the destination side so we don't have to start the replication process from scratch?

[Vitaliy S.]

Hello Tim,

As far as I got you right, you should create the replication job from scratch, because Veeam creates extra files for storing and tracking current state of the replicated VM as well as the available rollbacks (points in time you could restore to).
However, you may use replica seeding option which can be found at advanced job settings, this will create records for your VMs in Veeam SQL configuration DB and will move your VMs to a portable storage device, so after you could migrate VMs to the remote site and continue replicating using change block tracking feature, transmitting only changes over the leased line.

Thank you!

[tim.vogel@xtium.com]

We already have replicas at the remote site, and flying someone to the source site to move them AGAIN with a portable storage device is not really a good option.

There is no way to start the job so everything is created, cancel it, and then copy the vmdks in from the current replica? It can't do a differential initial (without CBT) job to verify the differences?

[Gostev]

Tim, no unfortunately this is not possible. Due to the product's architecture, seeding must be performed with the files produced by Veeam Backup during initial job run. This process produces replica VMDK files, as well as metadata file which content is directly dependant on the content of replica VMDK files produced during initial run. You cannot use this metadata file to start replication with any other VMDKs.