I have a new DPM 2016 environment (running UR1 on Server 2016 Datacenter) that has an MBS pool running on 1TB VHDX files via Hyper-V passthrough. The setup has been working mostly well, except for with protecting a large, deduplicated datasource. The volume being backed-up is a 15TB disk with 6.4TB used (with a 50% dedup savings) on a Server 2012 R2 box. The replica creation is currently at 132 hours and has synchronized just shy of 500GB.
The DPMRACurr.errlog on the DPM 2016 server is comprised almost entirely of this line, repeated over and over:
26581D6C12/1515:25:56.71418fsutils.cpp(4064)353C7761-E6C5-45E9-BF38-0E1FEF47E30DWARNINGFailed: Hr: = [0x80070002] : Encountered Failure: : lVal : HRESULT_FROM_WIN32(dwError)
I have noticed on an occasion or two that the synchronization rate will skyrocket briefly, transferring large amounts of data before the speed drops back down. For reference on our DPM 2012 R2 environment this replica creation would have taken maybe a few days to complete.
Has anybody else seen anything like this in their environments?