Improving replication time - snapshot merges

[L33T]

Hi,

Running 2012 R2 on a DR Cluster with nimble storage at both ends, my replication is performed within veeam to the DR Cluster after backup jobs complete. Using CSV volumes as per normal but when the replication job is running the snapshot merges take forever. I assumed that my storage network was quick enough, it`s a new CS1000 with SSD cache so no IO issues there... Would i benefit from going REFS and upgrading to Hyper-V 2016? Or is there any other tips to improving this? Reason for my question really is not just about the time it takes to merge but because sometimes they just time out then i end up with a duplicate VM or a failed job as veeam has timed out. I also wonder if the timeout values can be extended in veeam at all to compensate for the merge times?

Thanks all in advance.

[Rick.Vanover]

Hi there - REFS file system does introduce efficiencies for core Hyper-V operations. At Veeam we have promoted the REFS file system for backup storage, but there are indeed improvements for Hyper-V:

(From: https://docs.microsoft.com/en-us/window ... s-overview)

Accelerated VM operations - ReFS introduces new functionality specifically targeted to improve the performance of virtualized workloads:

Block cloning - Block cloning accelerates copy operations, enabling quick, low-impact VM checkpoint merge operations.
Sparse VDL - Sparse VDL allows ReFS to zero files rapidly, reducing the time needed to create fixed VHDs from 10s of minutes to mere seconds.

--

The Block Cloning should help with checkpoint merges - but with any change/recommendation, I recommend setting up a test environment on both ends if possible to reproduce these situations and see if the improvements will be noticed.

One other technology that may help is the Resilient Change Tracking with Windows Server 2016 also.