Any reasons not to enable deduplication in FSx for Windows?

Question

Setting aside single-AZ vs multi-AZ, we offer [two different pricing tiers][1] for FSx, those being with and without data deduplication.  Looking at storage costs, the pricing with deduplication enabled is half that of non-deduped FSx with the assumption of a typical 50% deduplication rate.  We've also documented that deduplication is transparent to the end user and [doesn't significantly affect the file system's performance][2].  With that said, are there any potential reasons customers should not enable deduplication?

[1]: https://aws.amazon.com/fsx/windows/pricing/
  [2]: https://docs.aws.amazon.com/fsx/latest/WindowsGuide/using-data-dedup.html

Accepted Answer

I would also keep these in mind:

There is a performance hit when dedup is enabled, so that should be kept in mind as well.

Workloads that might benefit from deduplication, but aren't always good candidates for deduplication. For example, the following workloads could work well with deduplication, but you should evaluate the benefits of deduplication first:

* General purpose Hyper-V hosts
* SQL servers
* Line-of-business (LOB) servers

https://docs.microsoft.com/en-us/windows-server/storage/data-deduplication/install-enable#enable-dedup

https://docs.microsoft.com/en-us/windows-server/storage/data-deduplication/interop

Any reasons not to enable deduplication in FSx for Windows?

関連するコンテンツ