I can!
Both. Now you can have synchronous mirror between 3 nodes. As a result you have still fully redundant and high performing (no cache flush on one node down) cluster if you take one node down for service.
And you have +1 async replication option (with either 2 or 3 node HA cluster) to store snapshot-based replica on. For DR purpose. So we do support 1, 1+1, 2, 2+1, 3 and 3+1 scenarios @ the same time.
Sure. Pretty interesting.
ReFS from MS is a joke. Bad joke.
Aitor_Ibarra wrote:Wow!
Can you expand a bit?
3 nodes HA - is that just like the current 2 node HA, but with a 3rd node, or is it handled differently (e.g. 2 node HA with async mirror to 3rd node?).
replication for deduplication - by replication you mean full 2 node HA for deduped targets?
I will save some RAM...
Also, now that Windows Server 8 beta is out, will you be looking for people to beta test how well it works with Starwind? (both as an initiator and a host OS). Will Starwind targets stored on the new file system offer any resiliency improvements compared with NTFS?