I have a Primary DPM server that is protecting various File servers and a Exchange DAG. Everything is working as it should on this server, but I am experiencing constant issues on the secondary DPM protecting these same data sources.
On the primary, two Protection Groups in particular are set up to protect two different volumes on the same file server. Volume D:\ is 46TB in size, with a deduplicated file size of 22TB (actual 39TB) whilst Volume E:\ is 25 TB in size, with a deduplicated file size of 2.5TB (actual 5.5TB).
Issue 1:
As expected Volume D: was initially replicated to the secondary DPM server at its undeduplicated size ( 39TB), as was Volume E:\ with initial replication to the secondary of 5.5 TB. So when is Microsoft going to support dedup on a secondary DPM ? It
seems daft to support dedup on the Primary DPM which is always more likely to be close to the original datasources on the LAN and not on the secondary DPM which is most likely to be placed offsite on a WAN !
Issue 2:
I have a big issue with the subsequent Synchronizations on the secondary server for E: which seems to transfer almost the full 5.5 TB every day ( I have it set to sync every 24hrs) - although the data is fairly static (i.e.unchanging) on that volume. On one occasion a Sync continued for over 48 hours for this volume and had transferred over 20GB (according to the DPM console) , until I manually cancelled the job - how is that possible on a volume with only 5.5 TB (UN-deduplicated) ?? What is going on here - has anyone any ideas?
Issue 3:
Another File server I am protecting on both the Primary & Secondary DPM server - always fails over to a consistency check on the secondary server - usually due to the fact it cannot access a particular file which results in an inconsistent replica . However the Sync (and subsequent restore point) on the primary DPM server from the original datasource is always completes without any issues. Again, has anyone any clues ?
I do get the impression that the whole Secondary DPM thing is not quite robust enough. I can only assume that as the Primary seems to protect the original datasources ok, that the issue is with the secondary reading the information on the primary DPM.