You should create every disk as a separate R0 and let ScaleIO handle failures.
Configure SSD's as write-through (no cache) and depending on if you have battery backed cache you can decide if you configure write caching or not for rotating disks.
This is exactly right, use R0 for the drives, WT (Write Through) for the SSD storage pool drives and WB (Write Back) for HDD if you have a DRAM protected Raid controller.
keep in mind you also need OS drives. usually you want to have these drives mirrored with Raid-1.
Carl-Adam_Lilje
32 Posts
394
0
Posted June 26th, 2015 01:00
You should create every disk as a separate R0 and let ScaleIO handle failures.
Configure SSD's as write-through (no cache) and depending on if you have battery backed cache you can decide if you configure write caching or not for rotating disks.
Carl-Adam Liljeholm
Sr. Speciality SE
EMC Software Defined Solutions