Backing Up Clones

GarryB

I am starting to deploy clones for real use in our development environment.

My question is when a full backup is taken of a clone it backs up the image size. Is there any suggestion on just how to backup the diff file created with a clone?

way0utwest

I don't think this is possible, but why would you want the diff disk of the clone? We don't have a way to apply this to another cloned image. Is there some use case I misunderstand here?

GarryB

The use case for us is going to be this data will be in QA for an extended period of time. Accounts will be aging and QA will be doing testing. If the data is lost all that would be lost as well.

I guess without being able to recover in Clone from a failure a that point would not work since that functionality is not there. And I guess I could recover from a SQL FULL and Diff. I just was not planning on that storage requirement as we move these processes down to lower level environments.

Just trying to look at all the risks and gotchas as we move this data around with SQL Clone.

way0utwest

The use case for clone is to quickly spin up new copies of a database. We haven't embraced the idea of a point in time clone that needs to be preserved.

If QA needs data for an extended period, our method would be to create a new image from that clone and use that. If you are worried about DR for QA data, as they make changes, then we aren't really set up to do that at this time.

Clone are meant to be short term in our use case. Create a clone, do something, drop it and restart. Or get multiple clones that can quickly be used to allow separate work in different environments. These aren't mean to be long term items that have a need for DR.

We do have the ability to automate a backup and a new image with PoSh, which can allow you to capture changes at a point in time for future clones, but if you intend for this to be a long term environment for QA, I'm not sure Clone is the right product.

Do you have needs for other copies? I'm wondering where your image is being used outside of QA.

GarryB

What we have done is take our production database create a cloned image and masked all PII data. We then take a backup of that and create another image that we then deploy to our Dev and QA environments. We have 4 dev and 4 QA environments and each get there own clone of the masked database image. It is just in QA where they are doing more elaborate testing where they need the data to mature for proper financial testing.

way0utwest

In that case, it's likely you need another image for QA periodically, if you think you'd need multiple clones. However, I assume the 4 QA environments will deviate and get different changes/data alterations over time. In that case, Clone doesn't necessarily help you here.

Is there a way to "age" out the data quickly? Maybe that's a way to deploy a new clone and update it over time. If not, I'd think that you might be better off with a native db backup/restore for QA.