I like to put extra 10-15% space on the result of this calculation due to the overhead of pool technical. Although pool technology would distribute IO on the device of a pool evenly, it also would break seq IO to Random IO and more CPU/RAM consumption as tradeoff.
It's really hard to tell from the information you gave. One factor is performance depends on how many disks this pool has extended to. I guess this LUN is not the only LUN in this pool, if that's the case, you won't get the full performance from all disks.
reseach
2 Intern
•
225 Posts
0
November 23rd, 2011 23:00
I like to put extra 10-15% space on the result of this calculation due to the overhead of pool technical. Although pool technology would distribute IO on the device of a pool evenly, it also would break seq IO to Random IO and more CPU/RAM consumption as tradeoff.
Eddy
SKT2
2 Intern
•
1.3K Posts
0
December 5th, 2011 13:00
200(6+2)= 25; what is the tdat size you have? is your tdev a single device or meta?
Jingyi1
2 Intern
•
199 Posts
0
December 15th, 2011 01:00
It's really hard to tell from the information you gave. One factor is performance depends on how many disks this pool has extended to. I guess this LUN is not the only LUN in this pool, if that's the case, you won't get the full performance from all disks.
mattcowger1
61 Posts
0
December 15th, 2011 12:00
It would be reasonable to test this using VMware's new IO Analyzer fling: http://labs.vmware.com/flings/io-analyzer
It should give you a reasonable idea of what your LUN(s) can do.