1 |
Am 04.05.2011 14:39, schrieb Florian Philipp: |
2 |
> Am 04.05.2011 11:08, schrieb Evgeny Bushkov: |
3 |
>> On 04.05.2011 11:54, Joost Roeleveld wrote: |
4 |
>>> On Wednesday 04 May 2011 10:07:58 Evgeny Bushkov wrote: |
5 |
>>>> On 04.05.2011 01:49, Florian Philipp wrote: |
6 |
>>>>> Am 03.05.2011 19:54, schrieb Evgeny Bushkov: |
7 |
>>>>>> Hi. |
8 |
>>>>>> How can I find out which is the parity disk in a RAID-4 soft array? I |
9 |
>>>>>> couldn't find that in the mdadm manual. I know that RAID-4 features a |
10 |
>>>>>> dedicated parity disk that is usually the bottleneck of the array, so |
11 |
>>>>>> that disk must be as fast as possible. It seems useful to employ a few |
12 |
>>>>>> slow disks with a relatively fast disk in such a RAID-4 array. |
13 |
>>>>>> |
14 |
>>>>>> Best regards, |
15 |
>>>>>> Bushkov E. |
16 |
>>>>> You are seriously considering a RAID4? You know, there is a reason why |
17 |
>>>>> it was superseded by RAID5. Given the way RAID4 operates, a first guess |
18 |
>>>>> for finding the parity disk in a running array would be the one with the |
19 |
>>>>> worst SMART data. It is the parity disk that dies the soonest. |
20 |
>>>>> |
21 |
>>>>> From looking at the source code it seems like the last specified disk is |
22 |
>>>>> parity. Disclaimer: I'm no kernel hacker and I have only inspected the |
23 |
>>>>> code, not tried to understand the whole MD subsystem. |
24 |
>>>>> |
25 |
>>>>> Regards, |
26 |
>>>>> Florian Philipp |
27 |
>>>> Thank you for answering... The reason I consider RAID-4 is a few |
28 |
>>>> sata/150 drives and a pair of sata II drives I've got. Let's look at |
29 |
>>>> the problem from the other side: I can create RAID-0(from sata II |
30 |
>>>> drives) and then add it to RAID-4 as the parity disk. It doesn't bother |
31 |
>>>> me if any disk from the RAID-0 fails, that wouldn't disrupt my RAID-4 |
32 |
>>>> array. For example: |
33 |
>>>> |
34 |
>>>> mdadm --create /dev/md1 --level=4 -n 3 -c 128 /dev/sdb1 /dev/sdc1 missing |
35 |
>>>> mdadm --create /dev/md2 --level=0 -n 2 -c 128 /dev/sda1 /dev/sdd1 |
36 |
>>>> mdadm /dev/md1 --add /dev/md2 |
37 |
>>>> |
38 |
>>>> livecd ~ # cat /proc/mdstat |
39 |
>>>> Personalities : [raid0] [raid1] [raid6] [raid5] [raid4] [raid10] |
40 |
>>>> md2 : active raid0 sdd1[1] sda1[0] |
41 |
>>>> 20969472 blocks super 1.2 128k chunks |
42 |
>>>> |
43 |
>>>> md1 : active raid4 md2[3] sdc1[1] sdb1[0] |
44 |
>>>> 20969216 blocks super 1.2 level 4, 128k chunk, algorithm 0 [3/2] [UU_] |
45 |
>>>> [========>............] recovery = 43.7% (4590464/10484608) finish=1.4min |
46 |
>>>> speed=69615K/sec |
47 |
>>>> |
48 |
>>>> That configuration works well, but I'm not sure if md1 is the parity |
49 |
>>>> disk here, that's why I asked. May be I'm wrong and RAID-5 is the only |
50 |
>>>> worth array, I'm just trying to consider all pros and cons here. |
51 |
>>>> |
52 |
>>>> Best regards, |
53 |
>>>> Bushkov E. |
54 |
>>> I only use RAID-0 (when I want performance and don't care about the data), |
55 |
>>> RAID-1 (for data I can't afford to loose) and RAID-5 (data I would like to |
56 |
>>> keep). I have never bothered with RAID-4. |
57 |
>>> |
58 |
> [...] |
59 |
>> |
60 |
>> I've run some tests with different chunk sizes, the fastest was |
61 |
>> raid-10(4 disks), raid-5(3 disks) was closely after. Raid-4(4 disks) was |
62 |
>> almost as fast as raid-5 so I don't see any sense to use it. |
63 |
>> |
64 |
>> Best regards, |
65 |
>> Bushkov E. |
66 |
>> |
67 |
>> |
68 |
>> |
69 |
> |
70 |
> When you have an array with uneven disk speeds, you might consider using |
71 |
> the --write-mostly option of mdadm: |
72 |
> -W, --write-mostly |
73 |
> subsequent devices lists in a --build, --create, or --add command |
74 |
> will be flagged as 'write-mostly'. This is valid for RAID1 only and |
75 |
> means that the 'md' driver will avoid reading from these devices if at |
76 |
> all possible. This can be useful if mirroring over a slow link. |
77 |
> |
78 |
> This should help in concurrent read and write operations because the |
79 |
> kernel will not dispatch read requests to a disk that is already having |
80 |
> trouble managing the write operations. |
81 |
> |
82 |
> On another point: Are you sure your disks have different speeds? SATA150 |
83 |
> and 300 are no reliable indicator because most HDDs cannot saturate the |
84 |
> SATA port anyway. dd is still the most reliable way to measure |
85 |
> sequential throughput. |
86 |
> |
87 |
> Regards, |
88 |
> Florian Philipp |
89 |
> |
90 |
|
91 |
`man 4 md` also states that the "the last of the active devices in the |
92 |
array" is the parity disk in a RAID4. |
93 |
|
94 |
Regards, |
95 |
Florian Philipp |