Friday, 30 December 2016

Part 2: Pair Replication Sites And Configuring Replication

In the earlier article, we saw how to deploy and configure vSphere Replication Appliance.

In this, we will see how to pair the replication sites and configure replication for a virtual machine. vSphere Replication 6.1 is managed via the vSphere Web Client. So, you will have to login to the web client with an administrator user. Now, in the production vCenter site you see there are no target sites for Replication. 


In this section, we will click "Connect to target site" option and you will see the below screen. Since my vCenters are in linked mode, I will choose "Connect to a local site" and select the DR site vCenter. If your vCenters were not in linked mode, then you will have to choose "Connect to a remote site" option and provide the PSC details of the remote site. 


Once the vCenter site is selected and configuration is done, you will see the Target Sites section being populated with the DR vCenter details.


Similarly in the DR site, you will see the production site vCenter in the Target Sites section.


Once this is done, we can proceed to configure replication for a VM. In my case, I will be choosing a VM called Router which has almost No data on its VMDK as it booting off a Floppy. This would be easier to complete as a test replication to ensure connectivity.

Right click the VM > All vSphere Replication Actions > Configure Replication.


We will be replicating from the Production vCenter to the DR vCenter, hence I will choose Replicate to a vCenter Server.


Select the DR site vCenter as the Target Site to send the replication data to.


Here we do not have additional replication servers deployed. Replication servers can be deployed to handle large replication load. If there are none, then the replication appliance will handle this traffic. So I will keep the default, Auto-assign Replication Server.


Select the datastore where the replicated data should reside in the Target Location section.


I will not check Quiescing or Network Compression. This is up to your requirement.


In this section, you will get to specify the RPO and Point In Time Copy for your replication of that specific VM. The more low the RPO the more frequent the replication is initiated and ensures all new data is constantly replicated. This would also mean the network would be under heavy load. 

Point In time instances mean, how many replicated instances have to be saved. If you say keep 5 instance for 1 day, then when a VM is restored from the replicated instance there will be 5 snapshots available and you can revert to any one of your requirement. After the 1 day mark those 5 instances will be removed and the next new 5 instances will be saved. 

More the instances to be saved, more the data space used on the destination datastore.


Once the replication is configured, the Initial Full Sync will start and you will see the below screen. 
Full Sync transfers all the VM data to the DR site. This can take some time depending on how big the source VM is. Post the full sync, we will be performing the incremental replication and the changes will be recorded in the persistent state file (.psf) 


Once the Full Sync completes, the Status will be OK and the replication details will be populated.


And now, if you browse the datastore which was configured to retain the replicated data, you will see the below files for the replicated VM.


Part 3: Recover a virtual machine using vSphere Replication.