1 |
On 24-Feb-14 7:27, Facundo Curti wrote: |
2 |
|
3 |
> n= number of disks |
4 |
> |
5 |
> reads: |
6 |
> raid1: n*2 |
7 |
> raid0: n*2 |
8 |
> |
9 |
> writes: |
10 |
> raid1: n |
11 |
> raid0: n*2 |
12 |
> |
13 |
> But, in real life, the reads from raid 0 doesn't work at all, because if |
14 |
> you use "chunk size" from 4k, and you need to read just 2kb (most binary |
15 |
> files, txt files, etc..). the read speed should be just of n. |
16 |
|
17 |
Definitely not true. Very rarely you need to read just one small file. |
18 |
Mostly you need many small files (i.e. compilation) or a few big files |
19 |
(i.e. database). I do not know what load you expect, but in my case |
20 |
raid0 (with SSD) gave me about twice the r/w speed on heavily-loaded |
21 |
virtualization platform with many virtual machines. And not only speed |
22 |
is higher, but also IOPS are splitted to two disks (nearly doubled). |
23 |
|
24 |
I did some testing with 2xSSD/512GB in raid1, 2xSSD/256GB in raid0 and |
25 |
3xSSD/256GB in raid5 (I used 840/pro SSD with quite good HW-controller |
26 |
but I think with mdadm it might be similar). Raid0 was way ahead of |
27 |
other two configurations in my case. |
28 |
|
29 |
Finally I went for 4xSSD/256GB in raid10 as I needed both speed and |
30 |
redundancy... |
31 |
|
32 |
Jarry |
33 |
|
34 |
-- |
35 |
_______________________________________________________________ |
36 |
This mailbox accepts e-mails only from selected mailing-lists! |
37 |
Everything else is considered to be spam and therefore deleted. |