1 |
On 03/01/2014 15:13, William Kenworthy wrote: |
2 |
> On 03/01/14 15:34, Alan McKinnon wrote: |
3 |
>> On 03/01/2014 09:25, Stefan G. Weichinger wrote: |
4 |
>>> Am 03.01.2014 07:52, schrieb Alan McKinnon: |
5 |
>>>> On 03/01/2014 00:46, Stefan G. Weichinger wrote: |
6 |
>>>>> BFQ only for the SSDs ? |
7 |
>>>> |
8 |
>>>> Yes. The scheduler knows how to deal with SSDs while keeping everything |
9 |
>>>> responsive even under load. |
10 |
>>>> |
11 |
>>>> BFQ seems a good fit for your workcase - desktop/laptop. For those, |
12 |
>>>> interactive performance is the most important thing. |
13 |
>>> |
14 |
>>> So you set BFQ for the SSDs and CFQ for the hdds ? I have both in my |
15 |
>>> desktop. |
16 |
>>> |
17 |
>>> |
18 |
>>> |
19 |
>>> |
20 |
>> |
21 |
>> BFQ for both is the recommendation. |
22 |
>> |
23 |
>> But do try it both ways to see how it performs and compare. |
24 |
>> |
25 |
> |
26 |
> hmm, is BFQ good for VM's too? I am currently using noops (storage is |
27 |
> ceph) and was going to experiment but have not had the time yet. |
28 |
|
29 |
|
30 |
I have no idea, but I'd like to find out. |
31 |
|
32 |
Instinct tells me one of the host or guest should be NOOP so that the |
33 |
other one can get on with scheduling without conflict. But I also reckon |
34 |
the question is waaaay more complex than that. |
35 |
|
36 |
I'd like to see the results of any benchmarks you do with BFQ on VMs |
37 |
|
38 |
|
39 |
-- |
40 |
Alan McKinnon |
41 |
alan.mckinnon@×××××.com |