1 |
Jesse, Rich wrote: |
2 |
> It depends. I'd like to see SW RAID keep up with dual 2GB fibre |
3 |
> attached to an IBM FAStT600. Then again, I haven't been given a budget |
4 |
> to test this with a Linux server. :) |
5 |
|
6 |
I have seen four FAStT600's spread across 20 x445s doing software RAID, and it is substantially faster than hardware RAID. |
7 |
Thanks, |
8 |
josh |
9 |
|
10 |
|
11 |
> |
12 |
> Also remember which RAID you're talking about. If you need parity, see |
13 |
> my previous post about BAARF. |
14 |
> |
15 |
> Finally, saying the CPU is not usually the bottleneck in a database |
16 |
> server is misleading at best. Being mainly an Oracle DBA these days, |
17 |
> faster CPUs can *cause* CPU contention. Like I said before, "it |
18 |
> depends". And like everything else, YMMV. |
19 |
> |
20 |
> Rich |
21 |
> |
22 |
> -----Original Message----- |
23 |
> From: Sean Cook [mailto:scook@×××××.net] |
24 |
> Sent: Wednesday, February 01, 2006 3:39 PM |
25 |
> To: gentoo-server@l.g.o |
26 |
> Subject: Re: [gentoo-server] RAID |
27 |
> |
28 |
> |
29 |
> -----BEGIN PGP SIGNED MESSAGE----- |
30 |
> Hash: SHA1 |
31 |
> |
32 |
> kashani wrote: |
33 |
> |
34 |
>>Sean Cook wrote: |
35 |
>> > Redhat In their manual: |
36 |
>> |
37 |
>> |
38 |
>>>"With today's fast CPUs, Software RAID performance can excel against |
39 |
>>>Hardware RAID." |
40 |
>>> |
41 |
>>>As I stated before, it depends on where your bottleneck is... if you |
42 |
>>>are not cpu bound, software raid is great! and will boost IO through |
43 |
>>>put on comparable hardware. If you are already CPU bound, forget |
44 |
>>>software raid, it will degrade your system to a crawl... |
45 |
>> |
46 |
>> |
47 |
>>Badly done tests circa 1998 without any sort of methodology, mention |
48 |
> |
49 |
> of |
50 |
> |
51 |
>>cluster sizes, etc is proof than any idiot can make a computer slower. |
52 |
>> |
53 |
>> I'll argue that a fully supported hard raid card is always |
54 |
> |
55 |
> superior |
56 |
> |
57 |
>>to a software raid by it's very nature, having local I/O cache and a |
58 |
>>dedicated chip. However there are definitely workloads where a |
59 |
> |
60 |
> software |
61 |
> |
62 |
>>raid is good enough that spending money on a hardware raid card is |
63 |
>>pointless. I can not imagine a case where all things being equal that |
64 |
>>software raid would be measurably faster. |
65 |
>> |
66 |
>> In the event that removing your RAID card makes your disk 5x |
67 |
> |
68 |
> faster |
69 |
> |
70 |
>>I'd also recommend removing the admin who setup the original system as |
71 |
>>well. :-) |
72 |
>> |
73 |
>>kashani |
74 |
> |
75 |
> |
76 |
> My post did say "Back in the Day" and it was around 1999 that we did |
77 |
> this however, then I was running unstable raid tools and they have come |
78 |
> a long way, and I was running 2.2 kernel. |
79 |
> |
80 |
> However, the only people I am aware of that say the performance is |
81 |
> better on hardware raid are hardware raid manufacturers and you... You |
82 |
> don't work for LSI do you? Most of the linux software folks agree with |
83 |
> me or take a mildly more conservative tone. |
84 |
> |
85 |
> eg: |
86 |
> |
87 |
> Redhat Enterprise Linux (circa 2005-2006) |
88 |
> http://www.redhat.com/docs/manuals/enterprise/RHEL-4-Manual/pdf/rhel-isa |
89 |
> -en.pdf |
90 |
> |
91 |
> Often the excess CPU power available for software RAID parity |
92 |
> calculations greatly exceeds the processing power present on a RAID |
93 |
> controller card. Therefore, some software RAID implementations |
94 |
> actually have the capability for higher performance than hardware RAID |
95 |
> implementations. |
96 |
> |
97 |
> |
98 |
> |
99 |
> Mysql: |
100 |
> |
101 |
> Hardware Versus Software |
102 |
> |
103 |
> Some operating systems can perform software RAID. Rather than buying a |
104 |
> dedicated RAID controller, the operating system's kernel splits the I/O |
105 |
> among multiple disks. Many users shy away from using these features |
106 |
> because they've long been considered slow or buggy. |
107 |
> |
108 |
> In reality, software RAID is quite stable and performs rather well. The |
109 |
> performance differences between hardware and software RAID tend not to |
110 |
> be significant until they're under quite a bit of load. For smaller and |
111 |
> medium-sized workloads, there's little discernible difference between |
112 |
> them. Yes, the server's CPU must do a bit more work when using software |
113 |
> RAID, but modern CPUs are so fast that the RAID operations consume a |
114 |
> small fraction of the available CPU time. And, as we stressed earlier, |
115 |
> the CPU is usually not the bottleneck in a database server anyway. |
116 |
> |
117 |
|
118 |
|
119 |
-- |
120 |
gentoo-server@g.o mailing list |