2012-02-15 05:43 AM
I've recently inherited a NetApp FAS3140 with 4 trays (14 disks per tray for a total of 56 disks) with the following setup:
22 disks for data
2 spare disks
2 parity disks
2 dparity disks
The 28 other disks have been configured with SyncMirror so there seems to be resiliency on top of resiliency! Also, I found the following link: which recommends to configure 2 spare drives per 100 in an array. This setup has 2 spares, plus the 4 parity/dparity disks, plus the 28 syncmirror disks!
I am unable to speak to the person who originally configured this array and therefore can't ask questions about why things are the way they are. I'm fairly new to NetApp so also still learning the product terminology and features...
So my questions are as follows:
1. Is this a rediculous amount of resiliency?
2. Does SyncMirror improve read performance of the array? If so, is it worth the cost of so much storage?
3. I was thinking of stopping the SyncMirror, reconfiguring the storage into 2 aggregates of RAID-DP, keeping 2 hot spares to reclaim some of the 60% worth of storage/disks dedicated to maintaining uptime. Thoughts?
Thanks for any help!
2012-02-15 05:50 AM
1. It depends on how valuable your data is.
2. I do not think so, at least by virtue of being SyncMirror. If you spread your data across the same number of not-mirrored disks, you likely achieve the same throughput.
3. It is up to you, really. There is no technical arguments either pro or con.