In the previous two posts of this series, I explained how to complete a seeding operation for a backup copy job in Veeam Cloud Connect, both for regular and encrypted backups. But Veeam Cloud Connect can also offer replication solutions to end users, so in this post I will explain you how to seed a replica job towards Cloud Connect.
Part 1 – how to seed a regular backup copy job
Part 2 – how to seed an encrypted backup copy job
Part 3 – how to seed a replication job (this post)
1. Seed the backup to the service provider
The creation of the seed is exactly the same operation described in Part 1. You may also want to use an encrypted backup to protect the dat during their travel from your premises to the service provider, but as the provider will need to import the backup into his own Veeam environment, at some point you will have to share the encryption password with them. The first operation is the creation of a local backup that contains the virtual machine(s) we want to replicate later towards the service provider.
Once the backup is created, send it to the service provider using a removable media.
Once the service provider has received the backups, the import process to make the replica VMs available to the customer is the same as the two previous examples. The provider simply loads the backup files into one of the tenant Cloud Repositories, and notifies the tenant that the backup files are ready to be used for the seeding.
NOTE: Veeam has published a knowledge base article, KB2158, that explains how to use replica mapping in this scenario. I honestly feel that the seeding option is way more simple, as the provider doesn’t have to do multiple replicas to load the virtual machine in his own environment.
2. Create the seeded replication job
This operation is done by the tenant, and to be successful both the service provider and the tenant need to be running at least Veeam Backup & Replication 9.0 Update 1 (build 220.127.116.111). In this version in fact Veeam added the option to use a cloud repository as a seed. The job has to be configured like this:
It’s important to enable the option for replica seeding. I usually also suggest to use network remapping to map local networks to the Cloud Connect networks easily. In the second step, I select my virtual machine called SQL. In the destination step, I select the Cloud Connect hardware plan, and the only datastore I have available is selected for me by the software:
After configuring the network mapping and the different storage options, I can also leverage WAN accelerator to further reduce the bandwidth consumption. Next, we arrive to the most important step of the wizard; in the Seeding step, I enable the Initial Seeding option, and I select the Cloud Repository exposed by Veeam Cloud Connect, where I know my service provider has uploaded my backup files. I don’t need to use Replica Mapping as my VM is not running yet at the service provider:
I complete the job creation with the Guest Processing and the Schedule options, and I immediately run the job for the first time.
3. The first seeded replication job
It’s interesting to see what happens during the first run. Immediately, the tenant can see something different in the job statistics:
The SQL virtual machine is ready to be replciated, but first the job is going to seed the virtual machine, using the cloud repository as a source that is local to Cloud Connect. I can easily see that data is coming from a local source because the upload bandwidth at the tenant I own is only 8 mbits, so it’s 1 MBps. The speed of the replication here is 29.5 MB/s, 30 times faster than what it would be when executed over the internet.
After the seed has been populated by the local replication happening at the service provider, the incremental run of the replication from the tenant to the service provider is also completed, and the statistics of this sub-job are again interesting. The virtual machine is 50 GB in size, and the seeding operation actually moved 18.4 GB from the repository. This would have been the same amount of data to be replicated over the internet without the seed. But thanks to the seed and the WAN acceleration, I had only 246 MB going over the wire, even if this was the first run, and the entire job lasted just an hour with also the seeding at the beginning:
After this short series of posts, you can clearly see as seeding operations in Veeam Cloud Connect can really help in many difficult onboarding situations.