0

sIs anyone using Hyper-V Production Checkpoints as an enhancement to Storage Replica? My thinking is that since Storage Replica is only crash consistent if we take hourly Production Checkpoints on the source server/cluster and keep them for 24 hours we will add the capability to fall back to a Production Checkpoint (or go back in time 24 hours) from our replica server, useful in a DR situation if the crash consistent copy turns out to be problematic. This would match the snapshot ability of Hyper-V Replica.

Has anyone done this, or read about anyone else doing this, in production?

EDIT: For clarity what I am talking about is the following:-

Server A is a Hyper-V host that has a main volume that has all the VM data stored on it including the checkpoint files. Server B has an identical volume which is kept synchronised with the main volume on Server A using Storage Replica. In order to provide for VSS application consistent recovery in the event of a DR situation a script is used to create Production Checkpoints say every hour on Server A. The same script deletes any checkpoint older than 24 hours, which will cause it to be merged in. In the event of a DR event where Server A is lost if Server B is brought online and the crash consistent Storage Replica of a particular VMs drives causes issues (say there is a database application with corruption causing DB not to mount etc.) then the Production Checkpoint can be used to restore the VM to an earlier time with a fully consistent VSS image.

The replication of checkpoints might significant increase the churn so real world testing would be needed to see how well this performs. Adding the Production Checkpoints to this Storage Replica scenario would match the features offered by Hyper-V Replica.

Has anyone tested this, done this, or read about any one doing this?

Improve this question

1 Answer 1

Reset to default
1

Production checkpoints are the default for new virtual machines starting with Windows Server 2016. You should use them.

Production checkpoints are "point in time" images of a virtual machine, which can be restored later on in a way that is completely supported for all production workloads.

It uses features supported by the guest to create the checkpoint, instead of using saved state.

Standard checkpoints capture the state, data, and hardware configuration of a running virtual machine checkpoints and are useful if you need to recreate a specific state or condition of a running virtual machine so that you can troubleshoot a problem, therefore more usable for development and testing scenarios.

Storage Replica enables replication of volumes between servers or clusters for disaster recovery. It also enables you to create stretch fail-over clusters that span two sites, with all nodes staying in sync.

As a conclusion, you should use both, but note that details depends a lot on the configuration and purpose.

Improve this answer
5
  • I'm sorry but I do not see how what you have written addresses my question. I was not suggesting using standard checkpoints. I was suggesting using hourly Production Checkpoints in order to get more value out of Storage Replica in a Hyper-V environment in a DR situation.
    – Christopher Edwards
    May 11, 2020 at 16:06
  • Storage replica is for Stroage. PCks are for quick-restoring your VM. What is better depends on your situation. Like do you have a large storage ?
    – Overmind
    May 12, 2020 at 5:12
  • I'm not asking what is better. I don't mean to be rude but have you actually read my question?
    – Christopher Edwards
    May 12, 2020 at 9:54
  • You are not actually asking something concrete there. Yes, it's tested, it works fine (for me, no performance issues whatsoever), but may not work fine in your situation/configuration. How well it performs depends on your hardware and load and you did not give any info on this.
    – Overmind
    May 12, 2020 at 12:28
  • Is "Has anyone tested this, done this, or read about any one doing this?" not concrete?
    – Christopher Edwards
    May 12, 2020 at 16:16

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .