2018-07-09 09:36 AM
New to NetApp and absorbing all the terminology and attempting to apply best practices to our new installation.
We have a FAS2650 with (12) X 960GB SSD's and (3) 212C disk shelves with (12) X 10TB drives.
I believe my question is simple, but challenged with finding an answer that directly relates.
I've read through various documentation and done the google drill, but not getting a clear direction.
With a trifecta mix of VMware ESXi hosts (VCenter), Linux hosts (think SuSE Enterprise) & an eclectic combo of Windows servers (2008 & 2012) ...
On one hand we've heard it best to create ONE large aggregate and then parse out these hosts on relevant volumes.
On the other hand, we've heard it recommended to create aggregates for EACH host platform.
Thoughts or sources on this inquiry?
Thank you in advance!
2018-07-09 10:50 AM - edited 2018-07-09 11:02 AM
In your particular case most probably you will be interested in the config with one big SSD aggregate attached to one controller, another big aggregate with NL-SAS drives connected to the second controller.
SSD aggr: 10 data + 1 parity + 1 hot spare
NL-SAS aggr: 2*(14 data + 3 parity) + 2 hot spare
Then apply QoS on volumes to separete hosts if needed.
Maybe, just maybe, you might be interested in taking 4 SSD from the first controller to add them to the second controller as read/write cache, but since FAS2600 systems already have read cache on board you might not need it.