Part:08 Zerto Checkpoints–Automatic and Manual
Checkpoints are very important part of Zerto’s technical innovation. It is strong selling point of the product and strong uses cases for Business critical application. In fact VMware vSphere replication quite recently added similar feature. These checkpoint provide point in time (PIT) copies of VMs. Point in time(PIT) allows VMs to be rolled back to any point of your request. PIT term is widely used in High-end Enterprise scale storage arrays
Checkpoints are automatically taken every seconds or as early as possible. These checkpoints are crash consistent and written to the journals by ZVM. During recovery you pickup crash consistent checkpoints in the journal and recover to this point
You can manually create checkpoints. Manually allows you to control time and date. You can think of it as a snapshot which you take it manually for particular reason.
So it gives you an option to recovery VMs to either manually or automatically created checkpoints allowing to do PIT recovery
Let’s see how to add manual checkpoint. Go to any tab of GUI you will see Checkpoint button there.
Click it, give name for the checkpoint. Select VPG for which you wish to take checkpoint and press Save.
Below is the event logged “insert tagged CP” in GUI and you should also see similar event in vCenter tasks
To recovery application using VSS agent
Zerto provides VSS agent. We discussed in this post here how to install VSS agent. VSS agent helps to backup application data in a consistent state. Application data consistent is of utmost importance when you want to protect and recover application data. All good backup product includes VSS integration. Though VSS only works with Microsoft products.
After you successfully install VSS agent shortcut by name “Add VSS Checkpoint” is created on the desktop. You can take checkpoint using this shortcut, command line or schedule task. This checkpoint is sent to ZVM. ZVM then add this to Journal. VSS ensures it is in application consistent state.
NB: Such checkpoints are initiated within VM but are considered for entire entire VPG. Other VMs in that VPG will have crash consistent checkpoint. It is design consideration of VPG as to how many application need a protection and if they have to be part of same VPG group as VSS can take application consistent backup of only one VM
Figure: 01 when you double click the shortcut, give name to checkpoint
Figure: 02 confirms checkpoint is taken successfully
Figure: 03 Checkpoint event is successfully logged in vSphere client
Figure: 04 Checkpoint event is successfully logged in GUI
In below figure you see the checkpoints which are available for you when initiated failover (Test/Actual). I have marked VSS endpoint as yellow. Both checkpoint has name on it. It helps you identify the reason for checkpoint
Finally few things to keep in mind regarding checkpoints
Checkpoints will have a performance impact. This impact is ruled by how much of the data is in memory of the VM. So be wary about this impact especially during production hours.
VSS Agent can take application consistent backup of single VM only, Even if VSS agent is installed on all VMs of a particular VPG group. This is due write order fidelity implemented by Zerto
Any changes in VPG that leads to re-synchronization of VPG will remove all checkpoints and synchronization starts from zero again. If there are no checkpoints there is no other way to get PIT (point in time).
Let discuss few things on checkpoints and Journal history. If you look at the slider here. This slide start position below is at 10:15:03 PM
And slider’s stop end position is at 10:56:28
It suggest you have checkpoints for last 41 minutes, 25 Seconds. But by default journal history is set for 4 hours, Ideally you should see difference between beginning and end position up to 4 hours.If it goes less than it e.g. if you have history for last 3 hours it starts sending warning
The slider shows a maximum of 180 checkpoints spread over the most recent 24 hours stored in the journal. The older the checkpoints over this period the fewer checkpoints are shown, with at least two shown per hour. The majority of the checkpoints cover the most recent hour in the journal. To be even more specific use the Manual Select option.
In Manual Select option you get more granular options to choose. It is checkpoint taken every 5 second if you observe carefully. It also suggest you have the option of rolling back to minimum 5 seconds. In above figure you can also see the checkpoint manually taken from the VM. This explains the value of this product. Such granular protection few years ago was possible using only costly enterprise grade storage arrays.
If you wish to follow entire series of Zerto go to the Landing Page