RAID 0 @ ATA/133/7200rpm's

[uvi]

Just wondering if you guys think this is even worth it. Or should i just stick my dual 40gigs together for the extra space. Seems like the most fun thing to do right now is a RAID 0 SATA RAPTOR config, which i intend on doing once i get the drives. However i do have these ata133's layin here, 40gigs each, and was wondering how much of a performance gain i might see using themin raid vs stand alones. And then if i get used to it, how much more of a performance gain i would see when i switch to raptors.

And while we're talkin raid, is it pretty easy to swap array types once you have one already setup? Lets say i decide to kill off my raid 0 array and want just typical dual drive system for extra storage space, will xp or some other software understand this and reassemble the data so that it is usable again? Or would i have to format and do it from scratch?..

Thx, peace out yaLLz!!

 

[djnes]

It depends on what your doing, but 99% of the uses will give no performance increases when comparing a RAID0 array and the same drives in a non-array format. Add the increased risk of data loss, and it's just not worth it. Raptors are nice drives and are a nice upgrade over any other drives on the market. I'm not sure what you mean by the most "fun" thing, but I broke my Raptor array and used the drives in 2 separate computers, and I lost nothing in performance. Even the high end computer makers are going back to shipping without arrays.

In terms of breaking an array, etc, all data would be lost, so it would involve a complete reformat to take regular drives and make them an array, and vice versa.

 

[Doctor X]

not true...


You have to do your homework and match your drives... There is specific things you have to do in order to set it up correctly.

Please see my posts here at this thread.

http://www.hardforum.com/showthread.php?t=810115

 

[djnes]

not true...

I certainly hope that wasn't directed at me. Please don't be yet another noob going around posting false info. Let's start with the facts:

Quite a few members of the forum, including me, have done pretty extensive testing. RAID0 is all but worthless, and absolutely increases your risk of data loss. Video editing is the only function we found that had any real difference in performance. Most people, me included, had their seek times raised by going to RAID0.

Second, raid arrays cannot be broken or assembled without re-formatting the drives. Simple fact....is now and has been a fact 7 years ago when I first started building arrays in servers.

Now, one final fact. It's been proven and beaten to death that RAID0 is a mostly useless myth built on nothing but hype. It's been debated, tested and agreed upon. It would be nice for everyone to catch up, and move on to something new to debate. RAID0 is left for dead.

 

[ambit]

RAID 0 is sooo 2001.
1 or 5 is the only way to fly

 

[djnes]

RAID 0 is sooo 2001.
1 or 5 is the only way to fly

I have a RAID1 array at home for my fileserver...complete peace of mind.

As far as RAID5 in a desktop setting, the only negative to it is the cost. Performance, redundancy, etc are perfect. I'm not sure about IDE, it's probably the same, but you need a minimum of 3 drives.

 

[Doctor X]

I certainly hope that wasn't directed at me. Please don't be yet another noob going around posting false info. Let's start with the facts:

damn... take a valum... dont go flaming other people. How in the hell did you take that wrong? I take offense that you are being a jerk like you was on the other thread to everyone.

If you read my posts.... you would see i am not a noob... I have been working with storage for many years. and I have done my own benchmarks. I was only refering to the performance misinformation only. I have seen up to 30% to 50% improvement depending on how i set up the raid0. If you set it up wrong, you can make things much worse than a single drive. I made this statement before:

I have been working with ide, eide, ata33/66/100/133, sata, and all the different flavors of scsi, SSA, and fiber channel and san (including the old fddi drives from long ago) for the last 12 years so I know a thing or two about this subject. I have used raid 0, 1 , 0+1,1+0 (yes there is a difference), 5, 50 (data striped across multiple raid5 arrays) and a couple of bastardized versions of the above (like 7and 70. Not even sure If i could explain that one again. The standard never took off due to cost.).

I know my raid and my storage and my servers and pc hardware. I work on anything from pcs up to $2 mil servers.
Read the posts i wrote before you explode again. And keep an open mind.

 

[djnes]

damn... take a valum... dont go flaming other people. How in the hell did you take that wrong? I take offense that you are being a jerk like you was on the other thread to everyone.

According to your profile, your a newb here...so let me explain one thing to you. Your reply to my original statement, is considered a flame. Now, if you don't want to be flamed back, don't throw the first stone. Second of all, I certainly wasn't being a jerk to anyone in that other thread, I was confirming the truthfulness of certain posts, while expressing my frustrations with the high level of misinformation being spouted here.

Second, if your choosing to bring this thread down to a low level, fine. If your going to flame me, I'd appreciate it if the grammar was readable...perhaps you meant "you were", instead of "you was". Now, about RAID, that truth is out there, as Mulder would say. As more and more sites, print mags, and individuals do the testing, RAID0 becomes less and less of an option. If you wish to continue to agree with my through flaming, be my guest. I'd choose another method of agreeing if you play to stick around for a long time.

 

[Steel Chicken]

Second, raid arrays cannot be broken or assembled without re-formatting the drives. Simple fact....is now and has been a fact 7 years ago when I first started building arrays in servers.

im pretty sure thats not true, ive done it with the PC in my sig
hell i took a raid array out of an old mobo, and put it in the new one (with different raid controller chips no less!!!) and it worked

 

[djnes]

im pretty sure thats not true, ive done it with the PC in my sig
hell i took a raid array out of an old mobo, and put it in the new one (with different raid controller chips no less!!!) and it worked

That's quite a bit different than turning drives in an array into single drives, etc. When the array is created, data is spread across the drives....regardless of the type of array. Some type of data is written across, even if it's just allocation table data and such. Breaking that array always requires a reformat of the drive....it's a basic principal of how RAID works as opposed to a non-array or a JBOD setup.

 

[uvi]

wow thanks, that definately answered everything i needed to know, and then some. Really appreciate the in depth answer.

 

[Steel Chicken]

That's quite a bit different than turning drives in an array into single drives, etc. When the array is created, data is spread across the drives....regardless of the type of array. Some type of data is written across, even if it's just allocation table data and such. Breaking that array always requires a reformat of the drive....it's a basic principal of how RAID works as opposed to a non-array or a JBOD setup.

well yeah if your gonna format and use the array as single drives, sure
but my point was this, you can go into the raid controller software bust the array, and either recreate it, or even put it in another machine, recreate it without losing data

but yeah, if you change types (go from spanning to non-spanning) and actually use it, of course youre gonna lose data

 

[Doctor X]

According to your profile, your a newb here...so let me explain one thing to you. Your reply to my original statement, is considered a flame. Now, if you don't want to be flamed back, don't throw the first stone. Second of all, I certainly wasn't being a jerk to anyone in that other thread, I was confirming the truthfulness of certain posts, while expressing my frustrations with the high level of misinformation being spouted here.

Second, if your choosing to bring this thread down to a low level, fine. If your going to flame me, I'd appreciate it if the grammar was readable...perhaps you meant "you were", instead of "you was". Now, about RAID, that truth is out there, as Mulder would say. As more and more sites, print mags, and individuals do the testing, RAID0 becomes less and less of an option. If you wish to continue to agree with my through flaming, be my guest. I'd choose another method of agreeing if you play to stick around for a long time.

Dont go slamming other's grammer if you dont know it your self :" I'd choose another method of agreeing if you play to stick around for a long time" If you would have read the entire post I was referring to, you would have noticed that there are significant caveats in order to set up a fast set. But instead you resorted to putting people down.


Besides ... back to topic. A common setup that greatly effects your throughput is the master/slave jumpers. It has been common knowledge that you do not want to set that to cable select. Well about 4 years ago that changed. For whatever reason I have not been able to determine, if you set up your drives master/slave, you get a performance penalty in a raid setup of any type.

I have read these reviews you are referring to, and with the exception of very few, their methodolgy is flawed. Usually, for the sake of time, corners are cut. In order to do thorough testing, plan on spending a couple of weeks dedicated time. On bigger arrays, one test run may last all day. Imaging drives introduces additional variables into the test, so you dont want to do it.

Another caveat that I confirmed this past weekend, while you can move drives around on the controllers and leave your raid0 set intact, performance will suffer if you do not put the drives back in the correct order. It made 5MB/s difference on my test.


Final thought, then I will say no more. Until you try my suggestions in a controlled environment, then please refrain from slamming me anymore. I am an noob only to this site. I just joined this forum. If you read my entire post from the other thread you would have known that.

For anyone else, if you would like to find out more information, do ask.

 

[djnes]

Until you try my suggestions in a controlled environment, then please refrain from slamming me anymore. I am an noob only to this site.

Done already. I stated this many many times over. I've done thousands of experiments throughout my college years adhering to the principals of the Scientific Method. I know how to do basic testing. All of my RAID examples were done with clean images in a controlled environment. Don't ASSume anything unless you know how I've administered my tests. My results have been replicated over and over by countless users on here....and scientifically, that begins to solidify them as fact.

Now, as a word of advice from a long time member here. Your getting caught in a lot of flame wars around the forum with several different people. Guess where the blame is pointed? Do what you want, but continuing your several flame wars back and forth with these people isn't going to make the mods happy. All of us longtime members are always looking for new friendly debates with a chance to learn something new. But when you argue against proven facts that people like GreNME and Phoenix86 post, your credibility is going downhill. Those two are extremely respected around here, and happen to be extremely knowledgeable as well.

 

[Doctor X]

I never read your reviews. I have maintained all along I have performed my own experiements. I never meant to get in a flame war with anyone. However, when someone challenges a statement that is contrary to what you believe in, instead of laying out the facts which I have done (with the offer of more information if asked), you and a couple of others attacked for no reason. I am sorry for even posting anything because everyone is acting like a bunch of spoiled kids. I read a comment if I wanted to have a discussion, then go to AIm.

I guess that newcomers to this board is not welcome, even with the herd mentality. I never flamed anyone, but i do take offense when someone gets nasty with me.

Let this be the end of it. No more flaming and I am calling a truce. But I will still be in the boards. If you disagree with me, do it in a grown up manner.

 

[djnes]

Here's the problem...and what your missing. We, as a forum, often take an issue, debate the hell out of it, and then agree on a final answer. We work often in conjunction with other sites to compare results. Many of us have full labs of hardware to do all kinds of testing. This forum represents the pinnacle of knowledge for DIYers. So, when you come strolling along, disagreeing with the proven facts, and giving some not so good reasons why you disagree, your not going to get much credibility. If you have all the knowledge and skills you claim....take part in the testing discussions and contribute your results. Don't wait until an issues has been proven, beyond a shadow of a doubt, to come in and start arguing about it. That would be the equivalent of starting a thread claiming the world is flat, and arguing with anyone who disagrees. That certainly isn't the way to earn respect and make allies. You forget we were all noobs to the forum at some point in time. It's what you do as a noob that builds your respect, or causes others to deny you. And as I said before, disagreeing with something the entire industry is finally agreeing to is not the way to go. It's a paradigm shift, because not too long ago, everyone was screaming how wonderful RAID0 was. But the paradigm has shifted, and I'm detecting your not too thrilled with being a little late. Also, as want to argue, keep in mind RAID performance is like video performance. Synthetic benchmarks don't tell the whole truth like real world usage and testing.

 

[Ice Czar]

not true....

Hardly a politic answer

Synthetic benchmarks don't tell the whole truth like real world usage and testing.

http://www.hardforum.com/showthread.php?t=810115
I used HD Tach and HD Speed.

while the assertion that "those sites"
http://faq.storagereview.com/SingleDriveVsRaid0

have a flawed methodology may in fact be true
http://www.tweakers.net/reviews/515

your methodology appears seriously flawed as well

Benchmarks
HDTach 2.xx

Strengths: Measures read and write STR. Presents a reasonable measurement of access time.

Weaknesses: STR is NOT representative of real-world performance. The STR measurements are taken from sample sectors staggered across the drive, often rendering results for RAID configurations useless. This is because this method often means a particular sample reading doesn't cross stripe boundaries, therefore producing results equal or similar to single-drive results.

HDSpeed
HD Speed measures both sustained and burst data transfer rates of your hard disks, cd/dvd-roms and floppy drives. It features a realtime graphical display.

at least all "those sites" where actually employing IPEAK SPT traces
http://storagereview.com/articles/200111/20011109Renaissance_1.html

my personal opinion is, that everyone is not qualifying their statements enough
broad generalizations almost always have exceptions, there are indeed a multitude of variables and access patterns, the unspoken and asumed arguement in djnes statements are "when compared to top of the line Raptors in a typical gaming rig" pressed he will immediately provide several caveates for video, graphics ect

while the good Doctor, has a rather brusque tone and is equally good at making unqualified and broad statements
in fact no "flaming" per se has occured, however, there is indeed a definate lack of cordiality

the point of the arguement here is, that lacking actual benchmarks able to definatively show a particular arrays multitude of variables out performing another specific drive
its a circular arguement from the "I know" percpective, and becomes largely a subjective call, which isnt really defensible, thus we site people that do have the neccessary benchmarks to definatively settle the arguement, but then they are accused of flawed methodology

last I checked IPEAK SPT ran about $900
then you can be accused of flawed methodology too

RAID 0: Hype or blessing? by Femme Taken Page 2

AT's and SR's benchmarks (1)

Before evaluating our own findings and having a look at the benchmarks conducted by AnandTech and Storage Review, we should determine to what type of applications our tests are related. The desktop environment not just implies spreadsheets, word processing, browsing, mail clients and other, rather 'light' applications; on the desktop of the power user p2p-sharing, image- and video editing and backing up all kinds of digital media are frequent activities, too. AnandTech, Storage Review and Tweakers.net all share the same audience: power users and tweakers who want to maximize power, speed and usability of their rig. Unlike the average secretary, our target audience is continuously multi-tasking, which can include downloading everything that can be found on the internet, while taking pictures at the highest possible settings. In short: they demand the highest performance from their desktop.

The testing methods used by all three sites have a few things in common. Storage Review, AnandTech and Tweakers.net all use the famous IPEAK Storage Performance Toolkit from Intel, that records disk access in Windows applications to a trace that can be played back as a benchmark on other disks or RAID arrays. Using the IPEAK toolkit, SR, AT and Tweakers.net developed their own benchmarks, which explains some of the differences in the results obtained. Sadly, AnandTech and StorageReview hardly provide any information regarding their methods, making it difficult to get an insight in the kind of tests that they performed.

> More

Page 4
Our observations clearly show that the testing methods of AnandTech and Storage Review are too limited to jump to conclusions about the use of striping in desktop environments - let alone that they could be used to sign RAID 0's death sentence. Both AnandTech's and Storage Review's results of the IPEAK largely contradict Tweakers.net's benchmarks, where RAID 0 does show significant improvements in the I/O performance.

In the case of Storage Review, this can be explained by their use of traces based on older software, which isn't representative of present-day 'power user' workloads. Using system benchmarks to assess storage performance is at the very least questionable anyway, because they mainly depend on CPU performance. The same thing goes for the loading times of games - not to mention the fact that a test limited to two games will never give a realistic view of performance in general.

Most striking is the fact that neither AnandTech nor Storage Review even bothered to test RAID in situations where the performance of the storage system really matters - for example when a backup application, anti-virus tool or compression program is active in the background while demanding content creation applications are running in the foreground. These are conditions that quickly give rise to noticeable latencies and in which single drive setups don't suffice anymore.

Tweakers.net bases its testing methods on intense system use of a 'power user'. Such users are, after all, our target audience. Besides traces of Business Winstone 2004, Business Winstone 2004 Multi-tasking Test and Multimedia Content Creation Winstone 2004, we created traces of several games using IPEAK Storage Performance Toolkit and we recorded disk access in I/O-intensive actions, such as copying files, defragmenting partitions, running a virus scanner and installing software. As is to be expected from a 'power user', we often employed multitasking, though not in an unrealistic way. All in all we created 27 tests, together resulting in our Desktop StorageMark 2004, Gaming StorageMark 2004 and Workstation StorageMark 2004 indexes. The result is a broadly oriented test suite, which gives a realistic image of performance in various situations.

in addition, for the general member reading this thread, your personal access pattern obviously comes into play, its entirely possible to configure single drives with a partitioning and file placement strategy that can give your access pattern a boost over dumping it all on a RAID 0 array

as Doctor X mentioned there are several variables in tuning a RAID 0 array
but then our access patterns often are quite unique
a review of >
As the Hard Disc Spins
the still incomplete Partitioning and Optimizing Tutorial
and this post
might give you some ideas along those lines

 

[Doctor X]

good info! Thanks... I will do some more reading.