I confess – I love NetApp® snapshotTM copies! They brought me to NetApp nearly 20 years ago, as I loved the effortless data protection they offer as well as the great people behind the technology. Stay tuned for my next blog where I cover the many features that snapshots enable, from short-term data protection to SnapMirror® replication for disaster recovery and SnapVault® backup, to our topic today, volume clones.
One of the drivers for NetApp volume clones was a software partner that ran hands-on training for their customers. Setting up database test environments only to tear them down after a week’s training was labor intensive, and the storage space was expensive. They loved snapshots like me and asked if we could offer a read/write snapshot. Engineering delivered FlexClone® volumes, and we added another snapshot-based tool to ONTAP® software.
Volume clones in Google Cloud NetApp Volumes
We’ve offered volume clones in NetApp Volumes since the start of the service. You’ll find it under Create a new volume from a snapshot. For the Standard, Premium, Extreme and Flex File service levels, these clones are full size, thick clones. They are created quickly as space-efficient thin clones, but the clone volume is then immediately split from the parent volume as a completely independent volume. This simplified the service and eliminated dependencies between parent and clone volume. But it eliminated one of the key benefits – saving storage space with thin clones!
For the Flex Unified service, we changed clone behavior. Instead of immediately splitting the clone, we leave it as a thin clone. Initially the clone volume consumes little additional space, and as you write to it the space usage will increase. This makes clones quick and inexpensive. But it does introduce the dependencies I mentioned before: you can’t delete the parent volume while thin clones exist. You also can’t delete the parent volume snapshot used for the thin clone. To delete the parent volume or snapshot, you’ll need to delete the thin clone volume(s) first.
Clone splitting
Splitting a clone volume from its parent is a low priority task, so it can take hours or even days to complete a split depending on the volume size. While the split is in progress, most snapshot operations are blocked. This means no replications, backups, additional clones, or even volume snapshot operations. This is why the Standard, Premium, Extreme and Flex File service levels don’t support replication on cloned volumes.
When Flex Unified was first introduced, the service only supported thin clones. Starting today, the service offers the ability to spit the thin clone volume on demand. This gives you the benefits of both – fast, inexpensive thin clones, as well as thick clones when you need them. Why might you need them? Perhaps you’ve made a clone for testing changes to the volume’s data, you are happy with the changes, and now you want to delete the parent and just use the clone going forward. Or perhaps both parent and clone volumes are very active. Over time the base snapshot used for the clone will begin consuming a lot of space in the parent volume. Splitting them lets them live separate lives with no dependencies.
Clone benefits
As mentioned earlier, our training partners find clone volumes useful for student lab environments. Clone the source volumes, then delete the clones when they are no longer needed.
A more common benefit is using clone volumes for testing. Perhaps you want to test a software change non-disruptively. Clone the volumes and make the change on the clones. It could be anything from a simple library change to a full software upgrade. Another common test area is for disaster recovery. Break your DR replication so the volumes are now writable. But minimize the time and cost to resume replication after testing by cloning the destination and doing your tests using the clone. Once the test is complete, delete the clone and resume the replication.
You can also use clones to distribute full copies of your production data files or databases to your developers. Avoid data corruption or performance impacts on the production data and speed up your development by giving them inexpensive, independent copies of the data. NetApp has thousands of customers who use this every day to accelerate their EDA or CI/CD pipelines. With Flex Unified thin clones and the new ability to split the clones when needed, now you have this superpower as well. And this clone army is completely loyal to you!
We've also updated the product documentation with a new section on clones: https://docs.cloud.google.com/netapp/volumes/docs/configure-and-use/volumes/manage-volume-clones