We use a std convention i.e. if the dev is below 1 TB we make a hyper of size 34 GB i.e. 38000 cylinders. Now if i want to create a device of 500 GB then i will need 500GB /34GB i.e. ard 15 devs of 34 GB each
Suppose i form a meta say AAA with 15 devs of 34 GB each and bind it to a pool say T1 : RAID-5 (7 + 1). Then how the device will be distributed.
Using Thin meta, the data will distributed in Thin Pool spreading workload over more spindles, same as using Thin Device. But each meach meta member gets a small percentage of cache, a striped meta is likely to offer more writable cache to the meta volume
Fenglin1
6 Operator
•
2074 Posts
779
0
Posted July 14th, 2013 18:00
Using Thin meta, the data will distributed in Thin Pool spreading workload over more spindles, same as using Thin Device. But each meach meta member gets a small percentage of cache, a striped meta is likely to offer more writable cache to the meta volume