2017-09-07 11:48 AM - edited 2017-09-07 01:54 PM
I have an aggregate that is being pretty hammered, and to gain a bit of performence back, the thought process is to change the cache portion of the aggregate from RAID-DP to RAID-4. Which is fine to do.
I am not sure if a aggregate snapshot would be able to preserve the changes that I am going to make, or will it preserve the RAID-Cache settings?
along with moving some of tghe workload to another aggregate, else we are looking on recommending a SSD disk shelf for the read write intensive applications.
Also I am not seeing the snap aggregate option within ONcommand, is this CLI only?
2017-09-08 02:18 AM
The process you are suggesting is coverd here: https://library.netapp.com/ecmdocs/ECMP1368404/html/GUID-C31EDDB7-7DDD-4042-995C-B93CF0F4B9DA.html
However, to confirm this configuration change will not be retained in an aggregate snapshot. Aggregate snapshots are now not set as standard since to preform a restore will restore all volumes in that aggretae back to that point in time in an inconsistent state. Not normally advisable, so all default installs now have the aggregate snapshot policy disable and reserve set to 0%.
There is also a minimum number of SSD dirves that is recommended being added to an existing Flash Pool aggregate, please see Section Expanding a Flash Pool Aggregate on page 13 of the Flash Pool Design and Implementation Guide http://www.netapp.com/us/media/tr-4070.pdf. A single SSD drive like an HDD may produce a hot drive and actually hinder performance.
If I recall correctly RAID4 SSD aggregates also have a maximum RAID group size of 8, not sure if you'll exceed that.
It is also strongly reocmmended to have one hot spare SSD in RAID4 configurations, not necessarily the case for RAID-DP, so just wanted to make sure this is true in your case...
An uncorrectable error in an SSD RAID group that is configured with RAID 4 and has experienced the failure of one SSD results in the entire Flash Pool aggregate being taken offline. It could also cause a loss of data that is cached in the write cache. Therefore, NetApp recommends using RAID-DP protection for SSD RAID groups and HDD RAID groups.
Hope this helps,