Our Cloud-DR product creates a replica of a VM in our London data centre (DR data centre), which can be used for disaster recovery. This guide will briefly discuss the product and its features.
Gere us a test VM running on our cloud platform, this is a live production VM.
This server has the Cloud-DR product running on it. We can launch a recovery test which will bring up the VM (without any network) at the DR data centre.
When we do the recovery we can select the snapshot to use, these are points in time where the data has been copied from Nottingham to DR data centre. This happens every 10 to 60 seconds, and the amount of recovery snapshots available is directly related to the amount of data the production VM is writing. For example, if a VM is writing 10's of MB per day there will likely be several days of recovery snapshots available, whereas if the VM is writing 10's of GBs per day there will likely only be a few hours of recovery snapshots available. This can change frequently, if the VM is being particularly busy the oldest snapshot will be more recent. How many snapshots you will have available can only be known once the product is live.
We select a recovery snapshot to use for the test.
Here we have started the recovery of the VM, as a test, to the DR data centre.
The recovery test takes a couple of minutes to complete - this time is dependant on the size of the VM, but generally is fairly quick.
During this time the replication is stopped, and the VM in DR data centre is powered down.
The recovery process continues.
And once completed the VM is powered up in DR data centre.
At this point the VM is still running as normal and is unaffected by the DR VM being online - during this time the DR date centre is not getting any updates from the live VM.
The DR VM can be logged in to and browsed. This is only possible by CWCS staff, we will not be able to provide you with access to the DR VM when a test is being conducted.
Failover
There are two options available for recovery:
Recovery of the live VM using a recovery snapshot from the DR VM.
This will overwrite the live VM with a recovery snapshot currently held on the DR data centre. All data written to the live VM since the time the recovery snapshot chosen will be lost. The live VM will simply revert to the time the snapshot was taken. Therefore, it is important that you have copied any data off you need to retain before the recovery starts; depending on the state of the live VM this may not be possible and data loss will occur.
Failover of the live VM to the DR data centre.
We will failover the live VM to DR date centre, reverting to the time of the snapshot chosen to restore. Once the VM is live on the DR date centre we will reconfigure our network to route all traffic destined for your IPs to route to the DR date centre, and then your VM.
Cavetas
We can not recover individual files from the DR date centre. While we can access the DR VM there is no connection to the VM to allow transferring data.
Recovery is dependant on the size of the VM.
The oldest recovery snapshot point is directly related to the amount of data written to the VM. When a VM is not writing much data there will generally be several days of recovery snapshots available. When the VM is writing a lot of data the oldest recovery snapshot point will be a few hours old.
