This message was deleted.
# microsoft-fslogixs
Slackbot
07/31/2019, 6:25 PMThis message was deleted.
r
Ryan Gallier
07/31/2019, 6:31 PMI don't do SOFS/CAFS, but i do front-end mine with DFSN
d
David Inglis
07/31/2019, 7:25 PMWe don't use SOFS, but you could front-end the cluster role name with a DFS namespace i guess if you wanted cross-site or cross-storage-array HA. So 2x completely separate MSFOCs, with their own separate storage and separate cluster role names, unified by a DFS name space. But .... then it's starting to sound like SOFS!
To me sounds overly-complex but then I am small-scale compared to some of you guys.
David Inglis
07/31/2019, 7:28 PMOr perhaps i've misunderstood the question that Leee may already have answered 🙂
t
Tim Riegler
07/31/2019, 8:09 PMThe idea was I have \\domain\share\fslogix as my DFS-N. Then, on the backend, I don't care what the actual server / share names are. I can build a brand new file server cluster, re-point the DFS-N to the new infra, and my clients don't have to change their config. It is a bit more complex, but not overly so (and is how we have our network configured today, without CAFS).
d
David Inglis
07/31/2019, 8:19 PMOK yeah i misunderstood 🙂 So standard DFS all good but in theory when traffic from client to VHD is redirected to a different back-end server behind the DFS-N (for whatever reason), there will be that drop in connection to the VHD - which is what CAFS gets around. However in real life - FSLogix should just enter it's retry process and reconnect to the VHDs after a momentary blip. End users might not even notice anything? CAFS = open-file resiliency at the the file share level, in this case FSLogix has it's own resiliency.
2 Views