Slashdot Mirror


Basics of RAID

Doggie Fizzle writes "RAID has been common in business environments for ages, and is now becoming more viable and popular for personal computers. This article focuses on the the basics of RAID, and spells things out for beginners or tech veterans. From the article: 'The benefits of RAID over a single drive system far outweigh the extra consideration required during installation. Losing data once due to hard drive failure may be all that is required to convince anyone that RAID is right for them, but why wait until that happens.'"

242 comments

  1. Slashdot: News for N00bs by Anonymous Coward · · Score: 0, Insightful

    Anyone ever get the feeling Zonk just doesn't get the Slashdot target demo?

    1. Re:Slashdot: News for N00bs by Anonymous Coward · · Score: 0

      Shit, I am a PHB, and I know what Raid is. Wouldn't be able to get rid of those damn carpenter ants without it.

  2. Bah we know this. Teach us how to raid panties! by Anonymous Coward · · Score: 0

    We're NERDS. We created raid!

  3. Holy Ads, bat-man! by temojen · · Score: 4, Insightful

    That's an awful lot of ads for a re-hash of well-known info. Are the editors sure this is frontpage worthy? It looks like a blatant attempt to get page views to me.

    1. Re:Holy Ads, bat-man! by trixy_1086 · · Score: 1

      What I find interesting is that a similar reply to yours by an AC got modded as a troll.

    2. Re:Holy Ads, bat-man! by Anonymous Coward · · Score: 1, Funny

      You think that's bad. See my article on why SATA is better than IDE. It's only more relevent now than when it was written, 12 years ago.

    3. Re:Holy Ads, bat-man! by Threni · · Score: 4, Informative

      Looks ok to me - not a single one. Want my AdBlock list?

    4. Re:Holy Ads, bat-man! by Nuttles1 · · Score: 3, Informative

      Ads...I see them once in a while, but they are quickly added to my adblock list. Really, should tech saavy /. readers see very many ads at all? I think the poster of the parent is a wanna be geek. Ok, enough sluphing off, back to work...

    5. Re:Holy Ads, bat-man! by Anonymous Coward · · Score: 0

      Yes, why does every little blog have to have its own advertising campaign to gain income. Is running a website that expensive these days?

    6. Re:Holy Ads, bat-man! by Eugene · · Score: 1

      what we need, is the ability for /.ers to rate the articles as well.. I'm getting tired to see POS articles showing up at /. the noise to signal ratio has become too much lately.

    7. Re:Holy Ads, bat-man! by SA+Stevens · · Score: 1

      I just rightclick on the image with Mozilla and never seen another graphic from that server again.

      Yes, with the 'light' setting of Slashdot config, you can turn off the graphics and miss barely a thing that matters on Slashdot.

    8. Re:Holy Ads, bat-man! by Spoukie · · Score: 1

      I didn't see any ads. This is the first time adbloc's (firefox extension)eating habits actually made feel like I was missing something!

    9. Re:Holy Ads, bat-man! by patmanDC · · Score: 3, Informative

      Forget all that. This is an excellent article on RAID.
      http://arstechnica.com/paedia/r/raid-1.html

    10. Re:Holy Ads, bat-man! by kgruscho · · Score: 1

      I've been wanting a system like launchcast, where everyone can rate everything. Then the system tracks people who rate similarily, mods up appropriately.

    11. Re:Holy Ads, bat-man! by SeventyBang · · Score: 2, Interesting



      The mystery becomes clearer. Hover your mouse over Doggie Fizzle's name above and see what resources Doggie is affiliated with. Then look at the uri of the "story".

      See any connections?

      (unfortunately, temojen didn't)

      Curiosity question: When did /. begin running an eyeballs++ service for those who want to increase the traffic to their web sites by submitting stories about their web sites?

      If it is an organized program, can someone get me the pricing for it?


    12. Re:Holy Ads, bat-man! by rthille · · Score: 1

      I've never bothered to setup an ad blocker, my mental filter seems to do a good enough job, and the fact that my browser loads them convinces the advertisers they should keep paying the website owners to produce content. Works for me :-)

      --
      Awesome furniture, accessories and cabinetry in Santa Rosa, CA: http://humanity-home.com/
    13. Re:Holy Ads, bat-man! by Thomas+DM · · Score: 1

      But.. it has even more ads than the site from the parent article! ;)

    14. Re:Holy Ads, bat-man! by CaffieneFiend · · Score: 1

      I am now shocked when an ad appears on my browser. Because of the adblock extension its like I am surfing the net prior to the 90's .com boom. Maybe that makes the ad that much more effective when it manages to get through.

  4. raid by Anonymous Coward · · Score: 4, Funny
    1. Re:raid by Pantero+Blanco · · Score: 1

      That's made even funnier by the fact that there's a can of it right next to my computer...It protects my data by ensuring that there aren't any roaches to crawl in any open expansion slots and fry my board.

    2. Re:raid by dgatwood · · Score: 2
      Roaches... no, but ants.... Have you ever seen what happens when a few hundred ants crawl inside a computer to get the crumbs that some 133+ computer repair guy dribbled inside it? :-D

      http://www.getipm.com/answers/computer-ants.htm

      --

      Check out my sci-fi/humor trilogy at PatriotsBooks.

    3. Re:RAID by geekoid · · Score: 1

      yes, but right now there is some 13 year old budding geek who now gets to learn about RAID.

      Just because your experienced, doesn't meen everyone else is.

      --
      The Kruger Dunning explains most post on /. http://en.wikipedia.org/wiki/Dunning%E2%80%93Kruger_effect
    4. Re:RAID by Anonymous Coward · · Score: 0

      A. Slashdot is not a site for newbies. It has never has been, and it never should be.

      B. There are thousands of sites for new users to learn about technology such as RAID.

      C. It's you're not, not your.

    5. Re:raid by Anonymous Coward · · Score: 0
  5. Excellent RAID reference by Logic+Bomb · · Score: 4, Informative

    There's an excellent guide to RAID levels (with pretty diagrams and such) at http://www.acnc.com/raid.html

    1. Re:Excellent RAID reference by cur3 · · Score: 1
      --
      how the end always is ...
    2. Re:Excellent RAID reference by Lawrence_Bird · · Score: 1

      There is also a good one at storage review:
      http://www.storagereview.com/guide2000/ref/hdd/per f/raid/index.html
      The original posted article is quite lame - they say raid 10 can only survive 1 drive failure, this is wrong it can survive 2 if they are not in the same mirror set.

  6. Or... by Joe5678 · · Score: 4, Insightful

    A source of information with far better content, that isn't simply an excuse to sell ads.

    Wikipedia

    1. Re:Or... by Anonymous Coward · · Score: 0

      Personally, I always thought that the arstechnica skinny on RAID was the definitive RAID article.

      http://arstechnica.com/paedia/r/raid-1.html

  7. Current HDD prices... by Manip · · Score: 3, Interesting

    Depending on the your budget here in the UK you can get an 80Gb HDD for around £35, so split over some time you should be able to afford two (or an extra one if you already have one). This is a good enough reason for anyone to try RAID.

    I myself currently have it setup to mirror my data across two 80Gb drives... Four months ago one of the hard disks died (funny buzzing sound, no access) but the manufacturers three year warranty was still valid, so I returned the drive to them for a free replacement. I received the replacement drive and shoved it in, mirrored the data back onto this new second drive and continued as before. If I hadn't have had this setup that data could have been permanently list. It also saves me from writing ten DVDs to store that much.

    1. Re:Current HDD prices... by Anonymous Coward · · Score: 0

      The problem is RAID controllers, not drive prices. Most on the market are garbage win-RAID.

    2. Re:Current HDD prices... by tool462 · · Score: 5, Insightful
      You may already know this, but it's worth mentioning to others who read this that may not. Your scheme works great to back up your data in the event of hardware failure, but in the event a virus or errant program corrupts the data, you'll still be wanting the DVD backups. That's because if a virus corrupts some files, it's doing the same thing on both drives, rendering the back up useless. RAID mirroring handles only one very specific type of data security. It's a very useful one, but it's important to understand the limitations or you can get bitten hard.

      Alternatively to DVD backups, you can also sync up your data on a regular basis to an external hard drive. This doesn't protect you if your house burns down, like DVD backups kept in a safety deposit box would do, but it does help you restore lost data after it gets corrupted.

      Ultimately, all these solutions require varying amounts of money, time, and effort, so you just have to decide what level of security you require and what you are willing to pay for it.

    3. Re:Current HDD prices... by ScrewMaster · · Score: 4, Insightful

      Ultimately, what it comes down to is that mirroring merely makes the hardware more reliable, it is not a backup technique.

      --
      The higher the technology, the sharper that two-edged sword.
    4. Re:Current HDD prices... by alienw · · Score: 0, Troll

      I've never heard of a virus corrupting data. I think that's a lot less common these days than hard drive failures.

    5. Re:Current HDD prices... by drsquare · · Score: 1

      Don't forget to include the costs of the RAID hardware to set up all these disks. 80Gb isn't a lot these days anyway, in fact it's only 10GB.

    6. Re:Current HDD prices... by morcheeba · · Score: 1

      Good point. Accidently delete or overwrite a file on a RAID array, and it's still just as hard to recover as if it were on a single drive.

    7. Re:Current HDD prices... by llamaxing · · Score: 1

      if you were interested in backup data, why not invest in an external hard drive? I mean, you won't have to worry about viruses screwing up the backup data, no DVDs necessary, and you can take it anywhere you want. Just remember, your friends can too, so make sure you invest in extra security options, too.

  8. Probably better by abrotman · · Score: 3, Informative
    1. Re:Probably better by Anonymous Coward · · Score: 1, Informative

      Arrrggh -- Redundant Array of Inexpensive Disks -- I know that there not necessarily inexpensive now days but that what RAID means -- before you go blabing about how its changed just tell me one thing!.... what the hell would a "dependent disk" be other then one thats already in a RAID???

    2. Re:Probably better by Joe5678 · · Score: 2, Informative

      what the hell would a "dependent disk" be other then one thats already in a RAID???

      A dependent disk would be one that can not be swaped out of the configuration. With independent disks you can remove one and replace it with a different one and go on about your way.

    3. Re:Probably better by obispo · · Score: 1

      Although the SIGMOD'88 paper meant "inexpensive" as part of the RAID acronym, early disk array implementations used specialized signaling to keep all disks in the array rotationally synchronized and even to have all actuators on the same cylinders at any given time, a la RAID-2 or -3. Disks in such arrays could be considered dependent on one another, unlike modern arrays in which head positioning is independent across disks. Obviously, an array of dependent disks is not adequate for all five layouts in the paper.

    4. Re:Probably better by zippthorne · · Score: 0

      Consumer hdds typically have 3 platters, since they are all on the same spindle and the heads are linked, maybe that's "dependant disks"?

      --
      Can you be Even More Awesome?!
    5. Re:Probably better by Solosoft · · Score: 1

      ooo lookie lookie ...

      Fancy info on your little debate.

      Should of read the website before you bitched :)
      But of course since you know it all you could also edit it too ...

  9. Give me RAID 5 by hobotron · · Score: 3


    Seriously, SATA hotswappable RAID 5, put an onboard controller on next gen motherboards, I dont care if its crappy compared to an expansion card, and you will have my money. Yeah we have RAID 0, 1 , 0+1, but no onboard commercial RAID 5 solution in mainstream motherboards. I know its more expenisve, but its also more efficient, and with every failed HD common users encounter the market gets bigger.

    --
    There is truth in humor.
    1. Re:Give me RAID 5 by Ann+Elk · · Score: 2, Informative

      Slightly OT, but this site is frequently updated with the current state of SATA/RAID support under Linux.

    2. Re:Give me RAID 5 by kayak334 · · Score: 3, Informative

      I've been using this board from Asus for about 6 months with onboard SATA RAID5. It cost $120 from Newegg.com when I got it, if my memory serves me.

    3. Re:Give me RAID 5 by gid13 · · Score: 1

      Personally, I'd be happy if the Linux kernel had a driver for my onboard Via SATA RAID controller, which I'm using for *gasp* RAID 0. Thing is, since Via has taken their sweet ass time providing drivers, I can't easily install Windows and Linux together on the RAID. Boo.

    4. Re:Give me RAID 5 by Nik13 · · Score: 2, Interesting

      In the last while we've seen just about everything get incorporated on motherboards: USB2, Firewire, multichannel digital audio, GB Ethernet and everything else. I'm hoping RAID5 will be part of the next features to be added to most boards too. There's really not much else I'd want besides that.

      There's always software RAID5 too. It would sound like it's slower, but I'm not 100% sure about that (less cpu load for hw raid is pretty much a given though). The other consideration is what happens in a controler failure situation. It might actually be easier to get the software RAID5 to work in a different computer (hw raid will need a new and identical controller to get to the data, which may take a while to get-if it's still available).

      --
      ///<sig />
    5. Re:Give me RAID 5 by TClevenger · · Score: 1

      No kidding. If I were a manufacturer of hard drives, I'd be _giving_ RAID-5 controller chips to motherboard manufacturers. Talk about increasing demand for your product.

    6. Re:Give me RAID 5 by InvalidError · · Score: 1

      I was thinking about doing something of that sort... but the specs for SATA PHY chips (so low-cost FPGAs without fancy IOs can be used) are somewhat scarce and most FPGAs with multi-gigabit transceivers cost over $200USD. (The elderly Xilinx XC2VP7 with 8xMGT currently costs ~$200 while the XC3S1000 which has 60% more logic capacity without the fancy IOs and PPC core costs $50.)

      A simple RAID solution can simply XOR the first two/four drive's data and store the result on the 3rd/5th drive. If any drive fails, XORing the remaining two/four drives will reproduce the failed drive's data so it can be returned or stored on the spare/replacement drive.

      The next step after the XOR RAID5 and 5+1 (hot spare) is to use Reed-Solomon (or equivalent) coding. By using the two extra drives (XOR/parity and spare) to store RS codes instead, the array can survive up to two failures until the failed drives are changed and sync'd.

      It would be a fun project if the SATA specs were more openly available and chips/PCBs more affordable. Well, I do not really have the spare change 5+ midrange SATA HDDs ATM anyway.

    7. Re:Give me RAID 5 by nmos · · Score: 3, Informative

      These days I think software raid is really the way to go, at least in comparison to the raid built into consumer grade raid cards. With software raid you should be able to move your disks to a working computer and boot up with a Knoppix CD and access your data if you have to. You can also raid individual partitions rather than entire disks. You could make a small non-raid boot partition on each disk which you sync regularly plus a larger data partition which becomes part of the raid.

      The above applies to Linux, I don't think the non-server editions of Windows can do anything but raid 0 (maybe raid1?). Possibly a BartsPE CD could be used to recover a failed Win raid array.

    8. Re:Give me RAID 5 by owlstead · · Score: 1

      Actually, there are many hybrid hardware / software raid solutions, where the xorring is done in software. Think the cheaper raid cards. These are not that much different performance wise from pure software solutions, but they may be easier to maintain.

      I myself would opt for a big tower, with software raid in a 4 or 6 bay swappable serial ata drive configuration. Raid 5 of course. 4 + 1 hot spare would be the optimum configuration probably.

    9. Re:Give me RAID 5 by afidel · · Score: 1

      Funny enough our HP ML100 with softare RAID5 on SATA is about 3x faster then our older ML530 with 12x 10K RPM SCSI drives in hardware RAID5 on a $2k controller when doing things like nightly backups. Unfortunatly the ML100 tends to buckle under heavy load and become ever more unresponsive, while the ML530 just keeps trodding along at its slower pace. Also you don't normally need an identical controller, just one in the same family. I know I have moved HDD's between Dell servers of different generations with different model PERC cards without issue.

      --
      There are 4 boxes to use in the defense of liberty: soap, ballot, jury, ammo. Use in that order. Starting now.
    10. Re:Give me RAID 5 by jabuzz · · Score: 1

      There is to the best of my knowledge no off the shelf SATA PHY chip available. Please correct me if I am wrong. Which is a shame because even with a moderately sized FPGA you could do a full hardware RAID5 SATA-SATA board that would kick ass. Think writes at full disk speed and much lower latencies than even hardware XOR assisted controllers can achieve. Oh and that last one is something people seem to forget about RAID, is that while through put can go up, latency does as well :(

    11. Re:Give me RAID 5 by InvalidError · · Score: 1

      Intel has specifications for SATA PHY interfaces, search for "SAPIS". Most of the SATA PHYs comply to this interface, though all the implementations of this that I have found are in the form of IP cores.

      SiliconImage had some (using their proprietary SATALite interface, press release here: http://www.siliconimage.com/news/press/detailpress release.aspx?id=134) but they disappeared from their product list earlier this year. Back then, the datasheets were only available on-demand. This particular chip was often used with older VIA chipsets and some PATA+SATA RAID boards.

      I think Marvell might also have some but all their SATA products appear to be on-demand.

      As for latency, it is only measurable when issuing read commands. In that case, the controller has to wait for the "slowest" drive and this is worst when using an array of disparate drives: active head positions and transfer speeds are uneven so data cannot be expected to be typically late by one rotation at most. I wonder how much of this an NCQ controller fitted with NCQ drives could hide.

      With all the gigabit serial links being deployed now, maybe next-gen low-cost FPGAs will start featuring some MGTs... they will need at least PCIE-x1 compliance sooner than later but probably would not feature more than four MGTs for a while.

      I wonder what the pricing on 4VFX20 is... it is the closest 90nm equivalent to the 130nm XC2VP12 and I would be curious to see how much above/below $200 it might be. (The fact that it is not yet available for open distribution screams hefty premium though.)

    12. Re:Give me RAID 5 by PSC · · Score: 1

      no onboard commercial RAID 5 solution in mainstream motherboards

      I don't think there's such a huge market for consumer RAID 5. RAID 1 offers safety against harddisk failure. Mainstream implies home users, and a RAID 1 (Mirroring) of two 400 GB SATA disks is more than sufficient for most people's MP3 and DivX collection.

      (BTW there is one mobo that supports up to eight SATA disks, four of which are configurable as a RAID 5. The other four only support RAID 0, 1, and 1+0. We evaluated the thing at work. I think it was the MSI K8N Neo2 Platinum for Athlon64 processors, but I'ld have to check to be sure. Interesting enough, on the manufracturers web page the board was said to have only four SATA ports.)

      I dont care if its crappy compared to an expansion card

      Well you should care. "Crappy" means not just bad performance, it means lack of reliability, and reliability is what RAID is all about. Accepting a crappy RAID controller means completely missing the point.

      If you do care about your data, get some proper RAID controller. (Fast if need to be, but that's quickly getting real expensive.) If, on the other hand, you don't care about your data, there's little point in RAID anyway.

      --
      --- The light at the end of the tunnel is probably a burning truck.
  10. RAID0 by fredistheking · · Score: 2, Informative

    RAID0 will increase your change of failure since you will loose all your data if a single drive fails. RAID0 isn't really redundant.

    1. Re:RAID0 by JonN · · Score: 1

      Yes, it mentions that in the article: RAID 0 is one of the configurations that does not provide redundancy, making it arguably not a true RAID array.

      --
      do.what.promptcmds
    2. Re:RAID0 by Anonymous Coward · · Score: 0, Funny

      Raid0: Kills data dead!

    3. Re:RAID0 by fredistheking · · Score: 1

      I was just pointing it out since the submitter makes it sound like RAID can be equated with redundancy.

    4. Re:RAID0 by joe_bruin · · Score: 2, Funny
      Yes, it mentions that in the article: RAID 0 is one of the configurations that does not provide redundancy, making it arguably not a true RAID array.

      Well, if it's a RAID array, I think it is redundant.
    5. Re:RAID0 by cagle_.25 · · Score: 1
      The array of disks is redundant. The data isn't. Seems like "RAID0" is an appropriate term for that.

      RAID0 isn't really appropriate for anything other than temporary files, anyway ... for exactly the reason you mention.

      --
      Human being (n.): A genetically human, genetically distinct, functioning organism.
    6. Re:RAID0 by Anonymous Coward · · Score: 0

      > loose all your data if a single drive fails.

      That doesn't make any sense. Your data will become not tight if a drive fails? What kind of strange slang is that?

  11. Um...Personal Use... by greymond · · Score: 1

    "now becoming more viable and popular for personal computers....Losing data once due to hard drive failure may be all that is required to convince anyone that RAID is right for them"

    I was thinking they were referring to "Joe Bob Home User" who is starting to use RAIDS more, which is true but, as far as I have seen they are NOT using it for RAID1 - (Mirroring and Duplexing) they ARE using it for RAID 0 (Striping) so their system apps/games run faster.

    1. Re:Um...Personal Use... by mink · · Score: 1

      Thats because the fucktards at MicroSoft (read on for why I consider them fucktards) have deemed all RAID levels above 0 as server only features.

      So for my desktop OS if I dont have a hardware RAID card outside of the OS, the only choice I get from MicroSoft even using the so called "Professional" editions of Windows is Raid 0. If I want better RIAD I have to use a server edition for the desktop.

      I have stopped hoping RAID in MicroSoft OS would go anywhere so I stopped buying Windows after the last time I sunk much $$ into a full version of Windows 2K Pro.

      --
      Well I've wrestled with reality for thirty five years doctor, and I'm happy to say I finally won out over it.
  12. Another helpful link by Toasty16 · · Score: 4, Funny

    Here is a link that explains the basics of computer hardware; I think that it's a good companion piece to the RAID article: http://www.angelfire.com/rings/judy_patch/

    1. Re:Another helpful link by Rufus211 · · Score: 1

      The sad part about this link is I can't tell if it's a parody or not. I've heard too many users say similar things.

    2. Re:Another helpful link by Anonymous Coward · · Score: 0

      You idiot - you gave the link but you didn't tell me how the hell back button works and what the hell is a CPU btw?

      What are Slashdotters now trying to screw me from my back? You bunch of idiots!

    3. Re:Another helpful link by Karma+Farmer · · Score: 1
      I *think* it's parody. But hell, I can't tell either.
      Every Hard drive has a unique IP address. An IP address is used by the internet so the governmnet can keep track of you. That way if someone is doing something illegal, such as hacking, the government or police can identify this person by their IP address, simmilar to a cars liscence plate.

      Some hackers however, have found ways to steal peoples IP address by using special virusses called trojan horses. Once the hacker has stolen someones IP address, they can use it while hacking to avoid being caught.

      stealing IP addresses is illegal and I will not be explaining how to do it here.
    4. Re:Another helpful link by Anonymous Coward · · Score: 0

      You Sir, are a moron.

    5. Re:Another helpful link by Barny · · Score: 1

      OMFG!!!1!! if you open paint and zoom in on teh immage you can read teh IP numbez, i r so going to use them!!111!! LOL

      --
      ...
      /me sighs
  13. RAID for "personal computers"? but why? by Rosco+P.+Coltrane · · Score: 4, Insightful

    Okay I guess it appeals to geeks and fancy computer modders and all. But really, when it comes down to it, a decent main hard-disk, a tray in the second bay for backup hard-disks, and a reasonable backup regimen that people keep up is all a "personal" computer user needs.

    Personally, I have 3 backup hard-disks, one that keeps a "clean" base system that I update every 6 months or so, and 2 that I do full differential backups on every 3 days. The "clean" hard-disk is kept off-site, and a script tells me when to do the backups on the other 2. And for very very important files, I just write them on a CD on the spot.

    With that, I've yet to lose a single file since I started using Linux in 93 or 94. My solution is cheap and doesn't involve fancy raiding. And I'm quite sure I overdo it, most people could do just fine with one main hard-disk, one backup hard-disk and a little discipline.

    --
    "A door is what a dog is perpetually on the wrong side of" - Ogden Nash
    1. Re:RAID for "personal computers"? but why? by DigiShaman · · Score: 1

      I use two SATA drives in RAID-0 for the simple purpose of bandwidth. It's a "must have" when working with 4GB+ AVI and MPG files for Adobe Premiere. Yes, I do video editing for music videos (for friends that need to promote).

      --
      Life is not for the lazy.
    2. Re:RAID for "personal computers"? but why? by doc+modulo · · Score: 1

      How about this.

      You do RAID 1 with 3 disks instead of 2. Two drives are for the RAID 1 setup and when it's time to do the backup you just power down or "prepare for hotswap", take out one of the RAID 1 drives which is now your newest backup and put the third drive back into the RAID.

      You wouldn't have to do "full differential backups every 3 days" anymore.

      Only think I've yet to look into, before I do this, is how to tell the RAID which of the drives it needs to update and which to mirror from. Wouldn't want it to overwrite all of my newest files with the versions of the old backup drive.

      --
      - -- Truth addict for life.
    3. Re:RAID for "personal computers"? but why? by Shawn+Parr · · Score: 1
      and a reasonable backup regimen
      Ah, but that is the rub, is it not? Many people are using mirrored arrays as an excuse to not backup.

      The university department I work for has a faculty member that teaches this through his actions. He claims everyone should backup (to cheap CDR's and DVD's mind you) and all of his workstations, and any students he has assisted with a purchase uses a Promise RAID1.

      Some of those machines have returned for assistance with issues, and I have seen some pretty odd hacks to get those RAID1 systems set up!

      This same faculty member has fought me on implementing a backup solution on the server that actual graded projects and finals exist on! Luckily I did convince them of deploying a RAID5 solution a while back, however as we all know that gives us uptime, but doesn't protect us when someone makes a boneheaded maneuver.

      We finally had a big enough issue come up that the Dean got us special funding for a backup solution (I had been begging for for 2 years). Well, my wife has new employment, and we are moving so I turned in my resignation. He has already gotten the funds frozen, likely for re-assignment, as apparently without me around they will not need it. It seems they only need something like that when someone tells them they do, and when they leave the department no longer needs it.

      Well, at least it has way better benefits than working in the real world...

  14. Still a single point of failure by L-Train8 · · Score: 5, Interesting

    With RAID, you still have a single point of failure. Instead of it being your hard drive, it is now your RAID controller. So what is the advantage?

    Since a RAID controller doesn't have moving parts, is it less likely than a hard drive to fail?

    --

    Don't forget that Friday is Hawaiian shirt day.
    1. Re:Still a single point of failure by Rick+Zeman · · Score: 3, Insightful

      With RAID, you still have a single point of failure. Instead of it being your hard drive, it is now your RAID controller. So what is the advantage?

      You get a new one under warranty or buy one...and your data is still there. If your drive dies and you get a new one your data's toast unless you have megabucks for Drive$aver$.

    2. Re:Still a single point of failure by Joe5678 · · Score: 1

      With RAID, you still have a single point of failure. Instead of it being your hard drive, it is now your RAID controller. So what is the advantage?

      For a home RAID system, the advantage is you can replace the RAID controller and still have all of your data. Unless of course the RAID controller corrupts all the data, but I'm not sure I've ever seen that happen.

    3. Re:Still a single point of failure by EvilMagnus · · Score: 2, Insightful

      Yes.

      The controller's still a point of failure. Indeed, with some RAID controllers if they go bad they corrupt data on *both* your disks, rendering both unusable.

      RAID protects against hardware failure of a drive.
      It does not protect against a bad controller or an OS snafu (for example, I once had the MSFT go bad on an NTFS volume, losing all data on a drive. RAID wouldn't have helped me there, either).

      So if you really care about your data, you should run RAID in conjunction with an off-disk backup solution. Preferably something that is regularly refreshed and kept seperately from your computer.

      --
      -EvilMagnus
    4. Re:Still a single point of failure by PhotoBoy · · Score: 3, Informative

      I've had my data corrupted by a dodgy controller... twice. I've been having terrible luck with the nvRAID provided on nVidia nForce 4 motherboards. Twice now the machine has locked up and on rebooting the RAID array is reported as damaged and a drive is missing from the array. A bit of Googling has revealed it's a common problem.

      Fiddling around in the BIOS disabling and reenabling RAID makes both disks show up again but putting them back into a RAID array seems to do no good as Windows always claims files are missing after doing this. If I reinstall Windows my data is always all still there in perfect condition, the hassle of reinstalling Windows and my apps is a pain though. So it's not totally corrupted, but enough to be a complete bitch.

      My feeling on RAID on the desktop is that it's a good idea but at least in nVidia's case it's being done on the cheap and is not totally stable. That said Intel's RAID controllers are superb and I'd use one anyday if it weren't for the vast amounts of heat and inferior performance of the P4.

    5. Re:Still a single point of failure by Anonymous Coward · · Score: 0

      Yes.

      Mechanical devices are always more prone to failure due to regular wear and tear. Where a tape drive might need to be retrofitted every few years, the electrical components behind it often run for years and years.

      Besides, even if the RAID controller dies, you do not lose the data. Just go get another RAID controller. The same cannot necessarily be said when the drive head comes smashing down on a platter and leaves a nice gash due to an accidental foot swing.

    6. Re:Still a single point of failure by kosmosik · · Score: 1

      > With RAID, you still have a single point of
      > failure. Instead of it being your hard drive,
      > it is now your RAID controller. So what is
      > the advantage?

      What kind of failure you mean? You mean getting your data damaged due to RAID controller failure? I've never heard of something like that but maybe it may happen. But I can bet that it will happen less than harddrive failure - so still having N drives and one controller is less likely to fail than having one hd - means RAID has advantage here...

      If you mean aviability - I think this is not the case here. We are talking about data safety. If your RAID controller fails - get a new one, plug it in and it should work. Or maybe get RRAIDC (Redundant Redundant Array Of Inexpensive Discs Controller). But then you will go to other place and find it is single point of failure, than another. :) It is not the case of having single point of failure (you will always find one) - it is the matter of probability of failure.

      Now for usual Slashdot bashing - somebody knows how often it is that data is lost (we are talking here about home users) not due hardware failure but software (like you know - Windows bug) one?

      > Since a RAID controller doesn't have
      > moving parts, is it less likely than a hard
      > drive to fail?

      Yes.

    7. Re:Still a single point of failure by diegocgteleline.es · · Score: 1

      Instead of it being your hard drive, it is now your RAID controller. So what is the advantage?

      Your RAID controller isn't who saves your data, so you don't lose your data, which is the point of several RAID modes?

    8. Re:Still a single point of failure by darkmeridian · · Score: 1

      That and when a RAID controller breaks, you can replace it with another one with down time but no data loss. When a hard drive breaks, you can replace it but you lose the data.

      --
      A NYC lawyer blogs. http://www.chuangblog.com/
    9. Re:Still a single point of failure by PhunkySchtuff · · Score: 2, Informative

      Yes and No.
      For a basic RAID, ie a host card in a computer attached to a bunch of disks, then yes, the card is the single point of failure. This is less of a problem than losing a disk because: the card is less likely to fail (as you guessed, no moving parts) and failure of the card doesn't necessarily mean loss of any data. Failure of the RAID card will mean access to the attached RAID array becomes difficult =) however some machines even have hot-swap adapter cards, in which case you swap out the card for a new one and you're back up and running. No downtime, no loss of data.

      For the usual kinds of RAID units used in enterprise, they're a hardware RAID enclosure with it's own RAID controller, attached via SCSI or FibreChannel to a host. Some of these units actually have two (or more) RAID controllers in them, each with their own FibreChannel connections (two or more, per controller) and if one controller dies, the other one takes up the slack (with a loss of performance, but not data). Going even further, you can have separate FiberChannel connections from each controller going through separate FC Switches, so even if a FC Switch dies, the SAN doesn't go down...

      What's the downside to all this? Big $$$.

      Cheers,
      Kai

    10. Re:Still a single point of failure by teslatug · · Score: 1

      It depends on what you call failure. If failure is losing data, then it's not a SPOF. If the failure is availability, then you wouldn't use RAID to save yourself. You'd use a high-availability cluster. When the controller fails, you fail over to another node with replicated data.

    11. Re:Still a single point of failure by drsmithy · · Score: 1
      With RAID, you still have a single point of failure. Instead of it being your hard drive, it is now your RAID controller.

      That's why you use software RAID for scenarios where you can't afford to keep a spare controller on the shelf, a decent warranty, or simply want the greater flexibility and reliability.

    12. Re:Still a single point of failure by ErikTheRed · · Score: 1
      With RAID, you still have a single point of failure. Instead of it being your hard drive, it is now your RAID controller. So what is the advantage?

      Since a RAID controller doesn't have moving parts, is it less likely than a hard drive to fail?
      A real RAID controller (not one of those crappy software-assisted RAID controllers; eg, anything under $400 or built onto a consumer-grade motherboard) is several orders of magnitude less likely to fail than your hard drives are. And even the best hard drives - server-grade SCSI for Fibre Channel - can be beaten to death within a year under an extremely demanding load (I had one database server that killed them in 6-9 months; this was before it was feasible to throw a few dozen gigabytes of RAM in a machine to keep the indecies in cache). Some of these controllers can cost thousands of dollars, and you most likely won't have them around the house. A few good ones for SATA / PATA can be had for between $500 and $1500 (see Areca, 3Ware, LSI Logic, etc.), and only people who really value their data will have these.www

      Do you need a 'real' RAID controller? The answer is simple: If you look at the price and your data is worth more than that, then the answer is 'yes'. Personally, I don't trust software-assisted RAID further than I can throw it.
      --

      Help save the critically endangered Blue Iguana
    13. Re:Still a single point of failure by Cyno · · Score: 1

      data redundancy.

      you don't get that with two power supplies or two keyboards.

    14. Re:Still a single point of failure by neilhu · · Score: 1

      There is another 'single' point of failure - the power supply. You may think that having a RAID cache, which is battery backed up may prevent this, but what about the HDD cache - every disk has write cache turned on by default - it's how disks get good performance.
      So, if there is a power failure, data in the HDD cache will never make it to the disk and since the RAID cache gets an acknowledgement from the HDD cache - the 8MB disk cache data disappears!

    15. Re:Still a single point of failure by zippthorne · · Score: 1

      Suppose you have two disks and "software assisted" motherboard RAID, Would it be better to hook them up as separate disks or as a raid then?

      --
      Can you be Even More Awesome?!
    16. Re:Still a single point of failure by ErikTheRed · · Score: 1
      Suppose you have two disks and "software assisted" motherboard RAID, Would it be better to hook them up as separate disks or as a raid then?
      By all means, RAID them if you'd like - but don't trust them with anything crucial unless you really understand the recovery process. And by understand, I mean do it once or twice with a configuration you don't mind using, until you are confident you know the quirks of getting back up and running - especially with Linux (and simulate failing BOTH disks!).

      Having said that, my main PC uses RAID-0 WD740 x2 (74GB 10K RPM SATA drives) for the boot disk / program storage. But I don't keep anything I would mind losing here (RAID-0 is living dangerously, even with these extremely high-quality drives - but damn is it fast). My data is all stored on an Areca SATA controller on my server (160GB x2 RAID-1 for boot/program, 300GB x4 RAID-5 (900GB net) for data.
      --

      Help save the critically endangered Blue Iguana
    17. Re:Still a single point of failure by cagle_.25 · · Score: 1
      It all depends on your needs. Several posters above have provided good links to RAID info; check them out.

      If you need storage space, just use independent disks. If you need speed for streaming data, use RAID 0 (but be aware that your speed will be severely curtailed by your software-assisted controller). If you need switch-over backup in case of disk failure (usually, servers need it, so that they can keep on carrying out their tasks), use RAID 1. (Do NOT use RAID 1 as a replacement for periodic backups of data.)

      If you need some combo of the above, use RAID 5 or higher.

      Each RAID configuration is trying to solve a particular problem ... so match your problem to the known solutions.

      --
      Human being (n.): A genetically human, genetically distinct, functioning organism.
    18. Re:Still a single point of failure by DigiShaman · · Score: 1

      For the record, I'm using the Promise FastTrak 378 controller on the motherboard. So far, I haven't had a single problem over a year and a half that I've used it. It's not the lastest and greatest in technology, but it's rock solid and does the job perfectly for my needs. Over all, I'm very pleased with it.

      --
      Life is not for the lazy.
    19. Re:Still a single point of failure by SanityInAnarchy · · Score: 2, Informative

      You do know that nvRAID is fake, right?

      As in, there's no RAID controller, it's software RAID done with BIOS code so that you don't have to dick with Windows as much.

      --
      Don't thank God, thank a doctor!
    20. Re:Still a single point of failure by steelfood · · Score: 1

      It really depends.

      The controller can mess things up. But, if it supports something like 51 or 15 (one is more redundant than the other BTW), especially over multiple controllers, the amount of redundancy in there probably could survive anything except for the physical destruction of several HDD's (four and up, depending on the number of mirrors).

      That's the beauty of RAID. It can be scaled to any level of redundancy necessary. RAID 1 (mirroring) can be performed multiple times across RAID 5. And, it can be done over separate controllers. And, with network speeds reaching the gigabit, I wouldn't be surprised if somebody offered a solution to do it over multiple privately networked machines, in multiple locations.

      The drawback to RAID (from a small business standpoint) is that to take full advantage of it, you need HDD's that offer exactly the same number of bytes of storage. That's usually the most difficult part. I'd assume that RAID controllers automatically scale to the HDD with the lowest capacity, thus wasting whatever's left over, and thus requiring the backup drives to hold the most bytes in the general capacity area (18GB, 36GB, 72GB, 100GB, 120GB, 180GB, 200GB, 250GB, etc.). Some people suggest buying from the same batch, but that actually is begging for problems, since RAID was not designed to recover from happening upon a defective batch in which all share the same defect (for example, the same sectors in every drive is weak or corrupted). For companies that can afford server farms, which probably have several layers of data redundancy (RAID, tape, backup servers, etc.), this isn't a problem, but for a person who only uses RAID for redundancy, well, you can imagine the horror of having lost everything.

      And to answer your question, RAID controllers won't fail logically (causing data corruption) from wear and tear, unless it is used in an extreme environment. It's like wondering whether your modem would eventually fail and corrupt your packets (the answer is not likely--I still have a 9600 baud ISA modem that is still chugging along on an old IBM AT). Usually, manufacturing defects or poor design result in bad controllers, and these are almost immediately noticeable. Now, a good and properly-configured controller and RAID setup should be able to take over a failed controller seamlessly. Unfortunately, having to replace controllers really defeats the true purpose of RAID, which is to be able to recover from hardware (specifically HDD) failures without powering down the machine.

      Oh BTW, RAID for homes is very basic, very watered-down. It's been relegated to little more than a fancy way to mirror HDD's. Very few home systems support hot-swapping, which means the machine will have to be brought down to replace the defective HDD anyway. The largest advantage comes in performance gains (taking about half the time to read for RAID 1, and half the time to write too for RAID 0). Most other RAID levels actually induce a performance decrease when it comes to writing.

      OK, enough ranting.

      --
      "If a nation expects to be ignorant and free in a state of civilization, it expects what never was and never will be."
    21. Re:Still a single point of failure by Karma+Farmer · · Score: 1

      Unless of course the RAID controller corrupts all the data, but I'm not sure I've ever seen that happen.

      It happens all the time. In certain circumstances, RAID can improve your uptime. It is never a substitute for backups.

    22. Re:Still a single point of failure by stmfreak · · Score: 1

      Disclaimer: I work in the storage industry for Isilon Systems.

      With RAID, you still have a single point of failure. Instead of it being your hard drive, it is now your RAID controller. So what is the advantage?

      Parent got modded up to Interesting? Okay, here is the clear advantage for RAID: While you may have a single point of failure such as a RAID controller or PSU or motherboard, your typically ghetto RAID (or JBOD) system still protects your data in the event of such a failure.

      You can replace a mobo or RAID controller and get your array back online, intact. Try doing that with the single failed hard drive approach.

      Since a RAID controller doesn't have moving parts, is it less likely than a hard drive to fail?

      Yes. Oh, yes. Verily, yes. Hard drives are timebombs. It's never a question of IF, but always WHEN.

      --
      These opinions guaranteed or your money back.
    23. Re:Still a single point of failure by PhotoBoy · · Score: 1

      I had suspected it was like that, given the way it needs its service running in the background and all the benchmarks that showed higher CPU usage than other controllers but I hadn't known for definite. Thanks for the info.

    24. Re:Still a single point of failure by ByteSlicer · · Score: 1

      Indeed. A few years back, I lost my hard drive (and pretty much everything else in my computer) due to a failing power supply. If I had been using RAID, I would have had two smoking hard drives instead of one. Backing up on a removable medium is a better solution imho.

    25. Re:Still a single point of failure by Barny · · Score: 1

      Highly recommend the highpoint 1640 card, it uses some (well, upto 10-20% at full load raid5) cpu time, but at least does mirroring and striping well on the card, with the added bonus that unlike a MB based raid solution, if the card dies (without killing the data) you don't have to buy a whole MB of the same type just to get your data back :)

      --
      ...
      /me sighs
    26. Re:Still a single point of failure by Salamander · · Score: 1
      Since a RAID controller doesn't have moving parts, is it less likely than a hard drive to fail?

      Yes, and that's one of three reasons RAID has an advantage. The second, as someone else has pointed out, is that with RAID (except RAID-0 which isn't really RAID anyway) you might lose access (temporary) but don't lose your data (permanent). The third is that the RAID controller is only a single point of failure for crappy RAID implementations. Host-based RAID, whether in software or hardware, just totally sucks rocks compared to external RAID, and even low-end external disk arrays nowadays support dual controllers with separate power supplies etc. Mid- and high-end arrays also have a large non-volatile cache for a huge performance win because your writes only need to wait for memory and can be "destaged" to the back-end disks in an optimal order at the controller's leisure (the larger the cache, the more you can optimize the back-end traffic). Typically you also get nice things like snapshots, replication support, LUN masking and mapping, etc. Anybody who has ever used a real disk array knows it's a whole different world compared to the way a home user thinks about storage. Yes, it can be a lot more of a pain to set up, and it's often hellishly expensive (an enterprise-level disk array will cost you far more than anything you connect to it), but the level of performance, functionality and robustness is just light years beyond what you get with dumb disks and host-based RAID.

      Disclaimer: I work in the storage industry, and have done so for quite a while. Make sure you don't confuse cause and effect, though. I don't say the technology's cool because I work with it; I work with it because it is cool.

      --
      Slashdot - News for Herds. Stuff that Splatters.
    27. Re:Still a single point of failure by jwhyche · · Score: 0


      Depends on how bad the controller dies. If it goes wonky and writes bad data all over your drives as it dies then you just as SOL as if your drives had died. Of course you powersupply can also blow up taking everything in the computer with it.


      That is what happened to me. Turn the fucker on, it made funny noise, then blew up. It was cool as hell, smoke, fire and all kinds of shit came blowing out the fan ports. Toasted everything on the MB but the graphics card, CPU, and the memory SIMM.

      --
      I read at +2. If your post doesn't reach that level I will not see or respond to it.
    28. Re:Still a single point of failure by jwhyche · · Score: 0

      Indeed, that is what happened to me and I had two drives RAIDed on that board. I wound up with two smoking harddrives instead of one. Moral of the story, don't buy cheap powersupplies. Besides, you will always have a single point of falure somewhere in the system. There is simply no way around it. Hell, you could raid your data cross two systems on ether side of the planet and still have it. One day your happly sorting your porn^H^H^H^Hdata, the next your an intersteller bypass. Hey, don't laugh. It could happen.

      --
      I read at +2. If your post doesn't reach that level I will not see or respond to it.
  15. Hit 'em when it hurts! by Gentlewhisper · · Score: 2, Insightful

    . Losing data once due to hard drive failure may be all that is required to convince anyone that RAID is right for them, but why wait until that happens.

    Because otherwise, you can tell them all about the wonders of RAID and all they'll do is just pretend to be interested while secreting thinking that you are some mad geek.

    Tell them about the wonders of RAID after they've been kicked in the nuts by a drive failure, and you sure as hell would be getting their whole undivided attention.

    Making the most of your effort man.. that's what it is :)

    1. Re:Hit 'em when it hurts! by Anonymous Coward · · Score: 0

      Pray tell, what does this have to do with the post you're replying to? Or were you just attaching yourself to one of the first posts in order to more effectively karma whore?

      That's what I thought.

  16. Money talks by cazbar · · Score: 2, Informative
    Losing data once due to hard drive failure may be all that is required to convince anyone that RAID is right for them, but why wait until that happens.

    Cuz the boss won't cough up the money until it happens.

    1. Re:Money talks by zenneth · · Score: 1


      Mod parent insightful.

      Let's just say there's this guy named Bob. Now, Bob has a couple of computers at his office. Some are almost modern: Pentium IIIs that originally had Win2k installed, but someone wiped 2000 and installed WinME. His receptionist's computer had an XP 1700+ with some Crucial memory... on a PC Chips board. His personal computer, in his office, was a Dell PII-266 with 64MB of RAM, running WinME. He claims that this computer is the "best in the office" because it mostly crashes when he's not using it, while it's just running the screensaver. He refuses to fix the receptionist's machine after the PC Chips board died, because he doesn't want to spend the $60 on a decent replacement board. This is the worst part of my job. Fixing other people's mistakes that were likely made simply because the boss was cheap and stupid.

      Someone removing Win2k and installing ME was the tops, though. Heh.

      --
      The Chronic *WHAT* les of Narnia!
  17. Article Summary by Anonymous Coward · · Score: 0

    RAID: Redundant Array of Independant (Inexpensive) Disks

    Why? If one disk dies you don't lose your data.

    RAID 0: better performance, worse reliability
    RAID 1: mirroring; no performance benefit
    RAID 5: striping w/ parity; perfomance benefit + reliability

    How? With Linux of course!!! 50f7w4r3 R41D is teh r0x3r

    Really guys, is this article really necessary? There's enough about RAID in basic CompTIA A+. I don't really think this needs to be posted on the front page. But what do I know; I'm just an AC.

    1. Re:Article Summary by Anonymous Coward · · Score: 0

      raid 1 gets you a performance benefit while reading. raid 5 has write deficits.

  18. Too expensive by pete19 · · Score: 2, Insightful

    As a (poor) student, I find that I simply can't afford an extra hard drive! I got a 2nd hand DVD burner from a friend for £15 and backup all my really important stuff (Code for university, photos, etc) every week. All my MP3s go on another DVD along with the hard disk, and they're "backed up" on my MP3 player anyway.

    As of yet I've never had a single hard disk failure... but I've not really got anything I'm bothered about losing, so RAID isn't worth it for me.

    --
    There is nothing more practical than a good abstract theory.
    1. Re:Too expensive by cagle_.25 · · Score: 1
      You're fine with that, and you're right: RAID isn't worth it for you.

      RAID isn't trying to solve your problems, anyway. Most RAID configurations are trying to provide higher rates of speed OR instantaneous fail-over solutions for servers. Not to worry!

      --
      Human being (n.): A genetically human, genetically distinct, functioning organism.
    2. Re:Too expensive by Anonymous Coward · · Score: 0

      for small ammounts of data, such as photos and code you can zip up the data, and email it to your own account at a free email service. As long as the email service is well known, and you remember to log in reguarly to keep it active, then you have an extra backup with no cost to you.

  19. SCSI RAID Yes, IDE RAID No by Anonymous Coward · · Score: 5, Funny

    IDE HDD Talking to IDE Controller:
    HDD: I'm gonna need more time for that write
    Contr: Yeah OK, go ahead good buddy
    Contr: What's up?
    Contr: What's up?
    Contr: Error: Drive controller timeout error

    SCSI HDD Talking to SCSI Controller:
    HDD: I'm gonna need more time for that write because I found a bad block
    Contr: Yeah OK, go ahead and remap that bad boy
    Contr: What's Up?
    HDD: Need more time to map that bad block
    Contr: Yeah OK, go ahead
    HDD: All done, grabbing the next command in the queue

    1. Re:SCSI RAID Yes, IDE RAID No by Anonymous Coward · · Score: 1, Informative

      I should probably add that SATA is closer in terms of functionality to SCSI than PATA. This does NOT mean that all controllers support it, however.

    2. Re:SCSI RAID Yes, IDE RAID No by Anonymous Coward · · Score: 0

      Non Clueful Quip

    3. Re:SCSI RAID Yes, IDE RAID No by drsmithy · · Score: 1

      IDE drives have had sector remapping for _at least_ 10 years now.

    4. Re:SCSI RAID Yes, IDE RAID No by NotBorg · · Score: 2, Informative

      Speaking of SCSI features, NCQ is now available on some SATA drives. It's not the full blown SCSI version, from what I gather, but it does bench well. See Tom's article on it.

      I have two RAID controllers populated with two drives each in RAID-0. One has two of Western Digital's Raptors (74GB 10,000 RPM). The other with two of Maxtor's DiamondMax HDDs(250GB 7200 RPM). The latter has NCQ and benches significantly better. Some of the difference may be related to one RAID controller being better than the other (I didn't benchmark both controllers with the same drives), but I suspect the bigger difference is due to the NCQ features of the DiamonMax HDD.

      --
      I want this account deleted.
    5. Re:SCSI RAID Yes, IDE RAID No by Jeff+DeMaagd · · Score: 1

      Better, maybe, but not better enough for personal/home server desktop. The same goes for RAID. It's nice, but expensive and doesn't benefit the individual enough to make it worth it.

    6. Re:SCSI RAID Yes, IDE RAID No by tchuladdiass · · Score: 1

      I don't quiet get the benifits of NCQ. From what I understand, it will re-order requests so that it can satisfy them in the most efficient manor, in otherwords, elevator head movement. But any decent OS should use elevator head movement anyways, so what's the point? Also, if the drive is re-ordering writes, then that can be dangerous, in that it makes things like journaling filesystems useless (unless the OS is smart enough to turn NCQ off while updating the journal).

    7. Re:SCSI RAID Yes, IDE RAID No by VoidWraith · · Score: 1

      It also has to do with the data density on the DiamondMaxes. Bigger drives always have better throughput. Seek time is probably worse on them though, since they spin slower too.

    8. Re:SCSI RAID Yes, IDE RAID No by Anonymous Coward · · Score: 1, Interesting

      The OS is supposed to use flush or ordering commands to make sure the drive knows when the order of writes must be preserved. Of course some IDE drives ignore these...

      Also, OSes only understand numeric block ordering, which doesn't always help with understanding which blocks are actually close together (think multiple disk platters). The drive can know where the disk heads and blocks actually are. Also, request queuing means drives can have the next few requests on board, instead of waiting for the next one to be delivered when the current one finishes.

    9. Re:SCSI RAID Yes, IDE RAID No by glazed · · Score: 1

      NCQ is all well and good, but a big advantage that Maxtor has is the drives with the 16MB cache. Having that extra cache memory seems to improve the NCQ performance a lot.

    10. Re:SCSI RAID Yes, IDE RAID No by Anonymous Coward · · Score: 0

      Yes that is true, however, the IDE drive does not communicate that to the controller, except once. And once that inital grace period is granted by the RAID controller, the controller/IDE drive do not have any further communication, so a timeout error will ocurr if the activity takes longer than expected. If it happens enough the drive may even be marked as failed (PROMISE RAID).

  20. Offset backups first, then RAID by inflex · · Score: 1

    Personally, I prefer daily backups to another HDD (use rsync, it's great), that way, if I make a major *oops* during the day I know I have a very recent backup immediately available, this is something that RAID cannot protect you from (the human failure).

    If then I've still got money to spare, I'll look at mirroring.

    http://www.mikerubel.org/computers/rsync_snapshots / is a great page to learn about using rsync to make easy backups.

  21. Fact checking... by KhaymanUCSD · · Score: 1

    Anyone else notice the error in the RAID 5 explanation?

    --
    Kneel before Sig!
    1. Re:Fact checking... by TrappedByMyself · · Score: 1

      You mean the extra parity block on the 3rd disk?.
      I'm really dissapointed that the author didn't explain the parity block. It's a great brain opener for someone never deeply exposed to boolean logic.

      --

      Help me take back Slashdot. When did 'News for Nerds' become 'FUD and Conspiracy Theories for Extremist Nutjobs'?
    2. Re:Fact checking... by KhaymanUCSD · · Score: 1

      Well there's that and on top of it the implication that is made is that you can only have 3-disk or 5-disk RAID 5 setups. In fact, you theoretically have a minimum of 3 and the maximum is based on what controller you have. I just feel if they're going to do an informative article, maybe checking the details might be for the best...

      --
      Kneel before Sig!
  22. Justification: by lanced · · Score: 1, Funny

    Yes!!!! I now can point my PHB to this article, and then to the article on how to turn IPods into a raid, and that should be enough to convince him to buy a few dozen. Now I will just need to explain to him why having a spare Ipod stored at my house will be a wise decision, you know, just in case the vendor closes it doors tomorrow or does something 'crazy' like switch architectures next year.

  23. There are two types of people: by caluml · · Score: 2, Insightful

    There are two types of people: Those that have lost data, and those that will.

    Don't forget, though kids - RAID won't protect you from deleting your own data, or a malformed script trashing stuff.

    1. Re:There are two types of people: by pete19 · · Score: 1

      RAID won't protect you from deleting your own data, or a malformed script trashing stuff.

      Which is exactly why I prefer to use a DVD backup instead. I get free DVDs from my university, so I keep one backup in my flat in case I need it quickly and one at a friend's place for off-site storage. That way I'm also (more) protected from fire. If my machine goes up in flames it doesn't matter how many HDDs I have, they'll all be damaged, but at least with DVDs I can save my really important stuff.

      --
      There is nothing more practical than a good abstract theory.
    2. Re:There are two types of people: by gunpowda · · Score: 1
      Data Loss in PCs:

      There are 10 types of people: oh, wait. What was I going to say?

  24. I for one welcome our new... by Anonymous Coward · · Score: 0

    ... eh Raiders.

    Seriously, can't we learn about how hot a redhead is in bed vs a blonde?

  25. Simple by Mensa+Babe · · Score: 1, Insightful

    "Losing data once due to hard drive failure may be all that is required to convince anyone that RAID is right for them, but why wait until that happens."

    Why? That's actually quite simple. Losing consciousness once due to a brick falling on the head may be all that is required to convince anyone that a helmet is right for them, but we don't see people wearing helmets, do we? Do people wear bulletproof vests? No? Why not? After all the benefits of a bulletproof vest over an ordinary vest far outweigh the extra consideration required during installation! Please, people, could we stop being such naive children thinking that our beloved industry is somehow special, that ones and zeroes have more value than humen lives? Could we please stop being so pathetic once in our lives? RAID may be a nice idea for "nerds that matter" but it is not for grandma. Period.

    --
    Karma: Positive (probably because of superiour intellect)
    1. Re:Simple by Anonymous Coward · · Score: 0
      actually grandma is the one that neeeds RAID 5. She's not going to backup her data EVER. and she's going to use the same computer for 10 years instead of buying a new one every 2 years so she will experience a hard drive failure. and after she losses all her data she'll probobly just stop using the computer.


      But if instead of that 160GB HD she has 3 40GB drives in a RAID 5 array when drive 1 fails she gets a dialog saying a drive has died, she calls up little jimmy from down the street to put in a new one, and then all is right with the world again.

      Technology is useless if it is not reliable

    2. Re:Simple by Anonymous Coward · · Score: 0

      Could we please stop being so pathetic once in our lives?

      Well, apparently you can't ...

      I'd almost think the parent was a karma whoring post, but it's so absurdly stupid that I can't imagine anyone would be tricked into moderating it up.

  26. Nice theory.... by dbc · · Score: 3, Informative

    ... but how often do personal backups actually happen? I'm one of those guys that has been taking home backups seriously for a long time, and has a collection of obsolete tape units to prove it. And backups still do not happen often enough if it requires me handling tape.

    Let's face it, discipline is a drag, that is why at work IT people are paid to schlepp around stacks of locked cases full of back up tapes to be shipped off site.

    So... for my home file server, I went to RAID mirroring, with a 3rd drive in a drawer. A mount-copy-umount chron job copies to the drawer-drive. Drawer-drive gets swapped and taken off site "when I think of it". Because... RAID only protects you from falling over hard drives. It does not proctect you from:

    1) Ooops, I wish I hadn't deleted that.
    2) Gack! My house just burned down! And took 10 years of tax data with it!
    3) Power supply goes wonky, causing both drives to scribble random scorfulentness everywhere.

    A home RAID system does not need to be expensive. Who needs hot swap? Use cheapo PATA drives. A few hours of down time for the wife and kids is OK. It doesn't take a big, bad CPU, and software RAID works great.

  27. Re:Give me RAID 5 AMEN by Anonymous Coward · · Score: 0

    from your mouth to Steve Jobs ears. If anyone will ever make RAID 5 a desktop standard its Apple. Its just too bad the G5 tower only has room for 2 drives. I'd take three 120GB drives in a Raid 5 array over a 300GB drive any day. People need speed and reliability much more than storage these days with the smallest drive you can get most places being 80GB.

  28. Simple by varmittang · · Score: 1

    Choose your Raid Click on which type of raid you want to know about, and it tells you what its good for, disadvantages and advantages.

    --
    -----BEGIN PGP SIGNATURE-----
    12345
    -----END PGP SIGNATURE-----
  29. Re:Arbitrary file inclusion! Yay! by Anonymous Coward · · Score: 0

    I've reported this in their forum. Hopefully somebody will fix it (or at least take the advice under consideration).

  30. is it possible to mix raid iwith non raid? by urbieta · · Score: 1

    Im thinking of adding a couple big disks to raid together for user files

    And keep my small hard drive for the operating system, if OS dies Id like to restore its image from the raid array.

    Am I dreaming here? or how can I do it? :D

    1. Re:is it possible to mix raid iwith non raid? by woah · · Score: 0, Redundant
      ... raid array.

      Isn't that a bit, you know... redundant.

  31. Windows RAID Over Firewire - Registry Setting by meehawl · · Score: 4, Interesting

    For those who have run out of internal space in their boxes, and who don't have external SATA or expensive hardware boxes, you can run RAID over Firewire.

    The problem, however, is that out of the box Windows refuses to "promote" an external disk to dynamic, which is required on all post-NT4 rigs for RAID.

    The solution is to add a semi-documented Registry flag, EnableDynamicConversionFor1394 .

    HOW TO: Convert an IEEE 1394 Disk Drive to a Dynamic Disk Drive in Windows XP

    Couple that with a cheap 4-bay firewire JBOD box and any spare old enclosures and you are set!

    I run 2TB in various RAID configs on my Windows server (main and near-line storage). Have done so since 2002. No problems with the external boxes. The support for external firewire RAID is a little gnarly in Windows 2000 - volume must be mounted as a named virtual directory and cannot be mounted as a letter drive. Later Windows give you both options.

    --

    Da Blog
    1. Re:Windows RAID Over Firewire - Registry Setting by fostware · · Score: 1

      Almost all disk utilities can't pick up dynamic disks, meaning if there's something on that disk you need - you're stuck. I think R-Studio can work with RAID members, including the software type. Dunno how well it works with FireWire connections though.

      Although... a decent case is a much better option...
      A quick Google shows this 11-bay box
      (http://www.coolermaster.com/index.php?LT=eng lish& Language_s=2&url_place=product&p_serial=RC-810&oth er_title=%2BRC-810%2BCM%20Stacker%20810)

      --
      "We know what happens to people who stay in the middle of the road. They get run over." - Aneurin Bevan
    2. Re:Windows RAID Over Firewire - Registry Setting by superpulpsicle · · Score: 1

      You people sure it's not an issue with the Raid5 registry flag instead of the 1394? Try this Tomshardware link.

      http://www.tomshardware.com/storage/20041119/

  32. RAID and Home PCs by agsharad · · Score: 2

    With hardware cost falling steeply, when will it become viable for home users to start having RAID-based PCs?

    All said and done, many of us do keep fairly important data on our home PCs. How many of us make an effort to back it up?

    --
    Warm regards,
    Sharad Agarwal
    AlcoHaul: We lift spirits!
  33. blah blah blah by Anonymous Coward · · Score: 0

    RE:"Losing data once due to hard drive failure may be all that is required to convince anyone that RAID is right for them, but why wait until that happens."

    data backup is what CD Burners are for, well they are plenty good enough for me.

  34. RAID 0+1 vs RAID 1+0 by georgewilliamherbert · · Score: 2, Informative
    This is a non-trivial question:

    When you're setting up a RAID set using both striping and mirroring, do you want to set up two stripes and then mirror between the stripes (0+1), or do you want to set up mirrored pairs and then stripe those mirrored pairs (1+0)?

    This is a quiz, and your data will grade you.

    What you want, by far, is RAID 10 (1+0).

    When you set up two stripes and then mirror across them, if you lose two disks, any disk in the first stripe and any disk in the second stripe, you lose all the data.

    If you stripe across mirrored pairs, then the only way to lose data is to lose both drives in one of the mirrored pairs. You can lose any other disk than the second drive in a pair, or even many more disks, as long as they aren't both in the same mirrored pairs.

    This doesn't make a difference with 4 drives. At 6 drives and up, use 10. Your data and users will thank you for it.

    1. Re:RAID 0+1 vs RAID 1+0 by cosmic_0x526179 · · Score: 2, Informative

      Keep one thing in mind about adding more drives.. Mean Time Between Failure falls as a function of how many drives you are using. If the drives are rated at 50,000 hours MTBF and you run a 2-drive mirror, you now have a MTBF of 25,000 hours on the array (not on a single drive). Likewise 4 drives would drop it to 12,500 hours. This is not a bad thing, just expect that with more drives spinning, the odds that one of them will woof is greater. Law of averages kinda thing.

      --
      This msg is brought to you by the letter 'W'.. for Worthless Wuss
    2. Re:RAID 0+1 vs RAID 1+0 by georgewilliamherbert · · Score: 1
      Keep one thing in mind about adding more drives.. Mean Time Between Failure falls as a function of how many drives you are using.
      Correct.

      Unless you need high performance (Database spindle count, for example), using as few drives as possible saves on expensive RAID controller or RAID unit chassis slots, and improves MTBF of the set of disks.

      Home users won't typically care, they want minimalist solutions, only talking about a few drives.

      Enterprise sysadmin / data managers care; when you get up into tens or hundreds of terabytes, even if your storage uses SATA drives underneath, you're talking about dozens or hundreds of drives. Then you get into 10 vs 01, hot spare plans, whether SW RAID on top and mirroring across RAID units makes sense, etc etc.

      Rule of thumb I have seen and heard repeated around the industry is to assume a 2% or higher failure rate for drives, per year. Even good, enterprise grade SCSI / FC drives. Probably should double or triple that for ATA / SATA components in large enterprise units (or small business or home units).

    3. Re:RAID 0+1 vs RAID 1+0 by Macka · · Score: 1


      In addition, with RAID 1+0 if you loose a drive and replace it, you only have a single disk to disk copy operation to wait for before the raid set is back to full strength.

      With RAID 0+1 you have to rebuild the stripe after a disk failure, and then suffer a stripe to stripe copy to restore the raid set, and that's slow !

      RAID 0+1 should be banned.

  35. We need new technology by zymano · · Score: 1

    We need something other than harddrives. They are the slowest part of computers and they crash.

    A laser/magneto drive would be more reliable and have more density ?

    1. Re:We need new technology by Armadni+General · · Score: 1

      We have something other than harddrives. It's called memdisk.

      And it's incredibly expensive.

  36. Here you go by kf6auf · · Score: 2, Informative

    I am running RAID 5 on my desktop server right here. It has a P4 3 year old Gigabyte motherboard. It's not hotswappable because it's not enterprise level (and I don't plan on having to hotswap all of the time, only when shit happens) but it gives me the RAID 5 that I like to use as a backup using software based RAID on Ubuntu Linux. After the install, it it would be just as easy for Grandma to use as if it were not RAIDed and I am certain any /.er could figure out the install for most any Linux distro.

    Can I have your money now?

  37. nt by Armadni+General · · Score: 1

    I'm planning a 800 GB (4x400GB) RAID 10 array whenever I come up with $1200. Now, I'm a total RAID newb, but even after raiding that articlem I still didn't learn anything. I'll take the Wikipedia article any day.

    1. Re:nt by drsmithy · · Score: 1
      I'm planning a 800 GB (4x400GB) RAID 10 array whenever I come up with $1200.

      Why not a 1.2TB RAID5 ? I'm guessing you're only setting it up for home or small office use, so it's not like you need the -10 for performance.

  38. So use two controllers by Sycraft-fu · · Score: 1

    Our most critical stuff is mirrored on seperate disks that are on seperate backplanes attached to seperate controllers. So the only single point of failure is the system, and if it goes down well the data isn't that useful anyhow.

    Also yes, RAID controllers are much less likely to fail. For the most part, if a solid-state device works for the first 30 days, it'll work forever if it's taken care of. There are exceptions, of course, but lacking any moving parts there just aren't a lot of ways for them to wear out, particularly for things that don't generate much heat and thus son't have much thermal expansion/contraction.

    With disks it not a question of if, but when they'll fail. Eventually, something will wear out and they'll stop working. Might be 20 years, but it'll happen.

    Unfortunately I don't have any empirical data so I'll just relate personal experience, even though it's not a proof: We see probably 10-20 harddrives fail per year for all our systems. Not unreasonable, we have a lot of systems. We see less than 1 harddrive controller fail per year. It almost never happens.

    1. Re:So use two controllers by C32 · · Score: 1

      Ever hear about electron drift?
      With todays feature sizes you'll be lucky to get 10 years of 24/7/365 use out of any reasonably complex integrated circuit...

    2. Re:So use two controllers by Wonko · · Score: 1

      With todays feature sizes you'll be lucky to get 10 years of 24/7/365 use out of any reasonably complex integrated circuit...

      Isn't 10 years close enough to forever for you? :p

    3. Re:So use two controllers by whoever57 · · Score: 1
      For the most part, if a solid-state device works for the first 30 days, it'll work forever if it's taken care of.
      You are quite wrong about this. Electronic devices have known failure mechanisms which determine the useful life of the device.

      IC designers are most concerned about electromigration these days. Electromigration causes the metal in the metal traces on chips to be moved from one location to another, causing thinner metal lines. Eventually this will cause the device to fail.

      Most mechanisms that affect lifetime are highly dependent upon operating temperature. Different applications areas (consumer, telecom, etc.) will have different goals as far as lifetime go, but 5 years is a typical minimum lifetime goal.

      --
      The real "Libtards" are the Libertarians!
    4. Re:So use two controllers by John+Hasler · · Score: 1

      > Ever hear about electron drift?

      I think you are trying to say "electromigration".

      Then there's the tin whisker and tin pest problems being brought to us care of the EU's new solder regulations...

      --
      Warning: this article may contain humor, sarcasm, parody, and perhaps even irony. Read at your own risk.
  39. Not until a HDD failed on me... by Soulfarmer · · Score: 2, Interesting

    I had my system hard drive fail fatally on me, emails and so forth, only some random backups elsewhere. Right then and there I decided that no more will a hdd failure steal my stuff from me and bought 4x120gb drives (size/price ratio at time was optimum) and a Promise controller. Now I got ca. 240gb RAID 01 setup, mirroring gives reduncancy and striping keeps the array at least as fast as those drives used separately.

    One hdd did fail on that array, and I just replaced it with warranty replacement hdd. No hassle, just carefree usage.

    The piece of mind is worth LOT more than those extra drives. I DO NOT like the menial job of building the OS from zero to working state, just because of a hardware failure, WHEN I can just as well avoid it.

    Proability of a failure greater than zero (0) is not zero. And I like it to be zero.

    --
    -Is the meaning of life vanity, or is vanity the meaning of life?
    1. Re:Not until a HDD failed on me... by Armadni+General · · Score: 1

      I'm actually a really big fan of RAID 10 instead of RAID 0+1. What made you go for 0+1?

    2. Re:Not until a HDD failed on me... by Soulfarmer · · Score: 1

      ummm.. not sure anymore. Not actually sure anymore which one I got. 01 or 10... with 4 disks, it doesn't really matter as long as only one drive fails before the array is rebuilt. Of course, if pair1 A and pair2 A drives fail at the same time, despair ensues, but if the two concurrently failing drives are A and B in their separate pairs, it is still rebuildable.

      But I really don't remember if mine was RAID10 or not. Performance-wise, I am not sure which is better even..

      --
      -Is the meaning of life vanity, or is vanity the meaning of life?
  40. RAID = Backup? by pipingguy · · Score: 1


    A lot of motherboards come equipped with RAID capabilities even if the end user doesn't know what the acromym means.

    Now that external USB/Firewire drives have become more affordable as a backup solution, I recommend that. Confession: I don't know WTF I am talking about.

    1. Re:RAID = Backup? by Spad · · Score: 1

      I don't know of any mainstream desktop boards that offer any more than 0/1/0+1, which is either unreliable (0) or inefficient (1/0+1).

      RAID 5 is what you want, it's just bloody expensive to get it (in hardware) at the moment. If one of your drives fails, you just take it out, stick in a replacement and it will rebuild the array for you from the parity information on the other drives. Fast read/write and good redundancy.

    2. Re:RAID = Backup? by Packet+Pusher · · Score: 1

      Until your raid controller fails and you realize that for no amount of money can you buy a replacement that will get your data back in any reasonable amount of time.

  41. Wow, a whole story I disagree with! Hmm.... by King_TJ · · Score: 1

    At one time, I thought RAID was teriffic. But honestly, I think it's way over-rated and exists primarily for the benefits of systems builders and manufacturers to sell people additional hardware and increase their profits.

    1. RAID implementations on most consumer-grade motherboards (EIDE RAID with Promise controllers on-board, and so on) are cheezy. I've tried using them for several years now, and I ran into lots of unexplainable "glitches" that never occured when I took RAID out of the equation. (EG. The RAID array would suddenly report a failed drive, yet when I'd pull the drive out and try reformatting/reusing it on another machine as a stand-alone C: drive, it would have no problems at all and S.M.A.R.T. reported it was fine too.)

    2. As others pointed out, the RAID controller is the new "single point of failure" - and amazingly, they do seem to go bad far more often than I'd expect for an expansion card on "better quality" RAID setups. I used to work for a place that slowly had every last one of their Dell "PERC II" RAID controllers die off, one by one, on their Poweredge servers - causing all kinds of hassle. (The card was no longer in production so finding identical matches for all the dead ones wasn't so quick or easy.)

    3. If you buy both of your drives at the same time, from the same place, they likely came from the same production run - so if one drive gives out, there's quite a good chance your other one(s) in your array will follow within a short period of time. After all, both have been running for the exact same number of hours of operation, at identical temperatures inside the system. More than once, I've seen a drive fail in a mirrored set, and then the spare died before the user got a chance to swap out the first one. So much for the data....

    4. If you actually get "hot swap" drive trays for your RAID system on a home-built PC, good luck with those too. Most of the ones I've purchased locally or over the net have been poor quality - creating yet another "point of failure". The cheap, little cooling fans in them usually wear out quickly, and I've had the little circuit/switch break on a couple that lets you put the key in and turn it to power the drive on/off to prepare it for removal/re-insertion.

    Honestly, I think the only solution is a good backup to off-site media if you have really important data to keep. For many home users, this probably isn't even a huge deal. Throw your resume and important documents on a flash drive or CD-R/DvD-R disc, and if the rest ever crashes - it's a good excuse to do a fresh, clean OS re-install anyway.

  42. Blu-Ray by Jambon · · Score: 1

    With Blu-Ray and HD-DVD on the horizon, each offering 30-50 GB per disc and in some cases more at a fraction of the cost, why would a normal consumer need to backup stuff using a RAID array? You're going to need to replace the broken hard drive anyway. I can see RAID being useful for businesses where they want as little downtime as possible, but is there really any point in buying more hard drives when you can get the same effect with DVDs or the upcoming BluRay and HD-DVDs? You can cite performance as an advantage using RAID 0, but as has been shown, there really isn't any.

  43. RAID is really here by JB72 · · Score: 2, Informative

    It's amazing how common RAID is now, especially (S)ATA RAID.

    In video editing, RAID is everything. External SATA RAID is the big thing now, and it works pretty well, even when it's OS based. What I haven't seen yet are (relatively) cheap SATA RAID 5 enclosures. That would be the Holy Grail of fast media storage.

  44. I didn't get any. by Mustang+Matt · · Score: 2, Informative

    Adblock must be doing it's job.

    --
    The man who trades freedom for security does not deserve nor will he ever receive either. - Benjamin Franklin
  45. RAID is just NOT reliable by Anonymous Coward · · Score: 1, Insightful

    I've sold Entreprise hardware / software / projects for over 13 years, and the handful of catastrophic data losses / service interruption / nightmarish server setup / stability issues my clients have had have ALWAYS been with RAID systems.

    Of course, if you're using RAID it's because you need capacity / speed / reliability, and while it seems the first two are indeed delivered, there is a HUGE reliability issue.

    Back in the days when i was selling Dell's own RAID array (can't remember the name), it was well known at Dell that the controller software was buggy, they fired the programmer but still sold the stuff. And thing haven't improved since. A couple of years ago it took a Compaq high-end VAR several MONTHS to get multi-gig RAID5 setup working. My brother works for a very large company, and after months of problems, their SUN lead person admitted that SUN has issues with their RAID controllers, especiallly the replacement ones / spare parts. Remember when SETI@HOME went down for a week a couple of years ago ? SUN RAID controler problems !

    My advice to clients is to avoid RAID whenever possible. Have spare servers, do very frequent backups, split your databases, have gobs of RAM... For the cost of a first-tier RAID system, you can buy a lot of other stuff, that WILL be useful in a lot more cases: fried servers, corrupted files... It's too bad the SCSI disks we use lag so much is capacity compared to ATA. I haven't done any SATA yet.

    The one thing less reliable than RAID is clusters, BUT i haven't handled clusters in a while, i'm talking 5 years ago. May be the new Linux ones are better ;-) I have no real-life experience with FC and the thing over ethernet yet. i'm VERY leery of my customers being used as beta-testers / guinea pigs.

    My 10 cents

    Sorry for the "AC", but since I'm naming names ...

    1. Re:RAID is just NOT reliable by Anonymous Coward · · Score: 0

      Guy I work with was at a site where some PHB thought that RAID and a SAN would be a great way to protect data.

      Vendor promptly supplied and configured the system to work.

      All was well.

      Then drives started dying left, rigth, and centre. Apparently there were firmware problems with the drives, and of course they all came from the same supplier/batch.

      RAID won't fix all your storage problems. You still need a good backup regime to protect your data.

    2. Re:RAID is just NOT reliable by mink · · Score: 1

      I've been managing, building, and servicing RAID setups for 10 years now and never see the kind of stuff you describe. That might be because the RAIDs I set up are SCSI, SSA, or FC.
      IBM RAID solutions just dont seem to suffer all the failures you talk about. Only things I have ever had happen was drives go bad and loss of a controler. Not once did any of these failures result in the loss of the entire raid array, or corruption of data. I'm not saying it cant happen, but it seems people are not maing good choices of hardware for RAID. All our external RAID setups use dual loop dual controller setups. Internal SCSI can not do this.

      --
      Well I've wrestled with reality for thirty five years doctor, and I'm happy to say I finally won out over it.
  46. how do failures behave? by mo · · Score: 1
    I've fiddled with raid a bit but haven't yet had enough failures under my belt to know how raid systems behave in real-world drive failures. Can somebody comment on their experiences for the following hardware configs:
    • What happens when, say, an IDE drive fails using software raid on linux? Does the machine stay alive, or do you need to remove the drive and reboot to get back to working?
    • How about failures with full hardware raid, like an LSI megaraid card. I've unplugged a drive with one of these and it beeps a lot, and seems to recover when they get plugged back in, but what about a true drive failure?
    • I'm also really curious about these new SATA raid controllers that libata calls "fake" raid controllers. Can they gracefully handle a disk crapping out?
    TIA to anyone who's seen enough failures to comment on this.
    1. Re:how do failures behave? by JayAEU · · Score: 3, Informative

      I've been using software-RAID with ATA drives on Linux for quite some time, so I can comment on the behaviour of an array containing a faulty drive.

      First off, let me emphasize how important it is to set up proper email notification (or pager etc.) for such cases! If you don't know about the failure, you're certain to get nice phonecalls from affected users.

      If you've set up the notification system (smartd and mdadm come to mind), you'll eventually get an email saying something like "Device: /dev/hdc, ATA error count increased from 0 to 1" and that it would be a good idea to check up on the hosts syslog.

      Checking up on the system, you'll find that the average system load has increased substantially, which is due to the system trying to persuade the disk to write to a faulty sector and the software RAID having to compensate, queuing the errors.

      Depending on how often the defective sector is tried to be written to, the load can increase to values of 10 and above, rendering the system unusable. This is a good time to halt it and replace the defective drive, partition it and "mdadm /dev/md0 -a /dev/hdc1" the new one into the array, starting to resync it right away.

      This may sound really horrible, but in practise it's usually less 60 minutes (counting from receipt of the first notification email) until normal operation can resume with such a system. This is assuming you have all spare parts stored somewhere on site.

      In genereal, I've found software RAID1 and software RAID5 on Linux to be exceptionally stable. I'm also very happy about the performance, given that all I'm using is a bunch of el-cheapo ATA disks. As for reliability, I'm convinced it can't be beat in the consumers' price range, since I've seen too many consumer grade hardware RAID controllers go down in a swirl when putting more than a light load on them.

      In the enterprise, I've seen companies move to software RAID on their Linux systems, because they found out that their only 5 years old enterprise hardware won't be getting any new spare parts anymore, which includes motherboards, CPU and IO controllers. Moving to software RAID on enterprise grade SCSI stuff allows them to move the entire system to another piece of hardware simply by moving the harddisk to it.

    2. Re:how do failures behave? by kentborg · · Score: 1

      On Linux...

      I like software raid 1 under Linux. I use it on my basement server. On day I was burning a CD image, and that is when one of my disks died. I never knew there was a problem, the CD turned out fine, the swap on raid 1 kept running, everything was happy, and the next morning I had an e-mail waiting for me telling me to get a new disk, which I did.

      The raid setup was as Redhat set up for me, but the cron script to watch for changes in /proc/mdstat and e-mailed me was homebrew.

      Unless you have a lot of money, stick with pure software raid. That way all parts are generic enough that you can be sure you can get replacements. If you want to do hardware raid make sure it is good stuff. I have heard stories of the fancy raid hardware having bugs. Make sure you have good live backups and hardware spares for everything unique. Use software raid, on the other hand, and you can get spare parts retail on a Sunday.

      Use raid 1 for your bootable volume, put your swap there too. If you need more space than you want to spend on raid 1, put additional disks in raid 5. I can't claim experience with raid 5, but under Linux it seems to have as good a reputation, and Linux raid 1 works great.

      I can't report on any experiences with SATA, but if you can reliably boot from SATA, Linux ought to be able to do you a nice raid 1 out of that. Be aware that SATA disks and motherboards are not yet commodity enough to sell you replacements on a Sunday. Doing SATA for hotplug backups might be nice, but the sweetspot for small systems themselves is still parallel ATA.

      You can do raid on top of floppies if you want--Linux is flexible. You could do encrypted raid. All pretty cool.

      A specific comment: Getting your lilo or grub set up so that you can boot from either drive is a bit tricky, and if the BIOS thinks a drive is good but it only half boots it will take human intervention to tell to BIOS to boot from the other. But the good news is that is you have a decent UPS you can just keep running until there is a person there to tend to the reboot. It is possible to have a set of grub choices where if a disk is slightly functional it might run well enough to boot off the second (working disk), that way even a naive human could tend the reboot (I've setup that). One could also setup a bootable CD that then offers choices for a real boot off of either disk (I haven't done that).

      -kb

  47. Build or Buy? by tabdelgawad · · Score: 1

    As a future project, I was thinking of building a RAID 5 solution from stock parts, but then I came across this

    http://www.lacie.com/products/product.htm?pid=1032 6

    Which seems not to be much more expensive than building your own. The same company has a nice line of other desktop and network drives (I have no connection to this company whatsoever).

    Drawbacks?

    --
    Imposing Libertarian views on everyone online since 1992.
  48. Tommorrow the basics of RDMS by jbplou · · Score: 1

    Seriously is this site about news or info you can find in books 15 years old?

  49. RAID is way overhyped by llZENll · · Score: 1

    Don't use stripping for performance unless you do video editing:
    http://www.anandtech.com/storage/showdoc.aspx?i=21 01&p=10

    Don't use RAID0 for redundancy because its not worth the effort, you still have a single point failure of the raid controller card. Stick with a 2nd device (usb drives are good) which you simply copy your important data to every week or so.

    Unless you can afford and need RAID5 (5 hard drives) forget about RAID and be happy your not wasting your time with it.

    1. Re:RAID is way overhyped by Anonymous Coward · · Score: 1, Informative

      1. RAID 0 is stripping. You are talking about RAID 1 and when your controller fails it doesn't take your OS and all your data with it.

      2. RAID 5 runs fine on 3 drives.

    2. Re:RAID is way overhyped by tomstdenis · · Score: 2, Interesting

      um ... do raid in software and be done with. That way the point of failure is either the kernel [roll back, quick fix] or the device [rush off and buy a new one].

      RAID-1 is a simple way to get a "reliable" store.

      Note that copying data to RAID [any of them] is *NOT* a backup solution. It's a "temp fix" for storing data.

      I use my RAID-1 [two 200GB disks I bought for 130$ each] as a simple "place to dump nightlies" which I then backup to CDR weekly. I do rely on the redundancy of RAID-1 in case I trash my CVS or home dir but my longterm backup strategy includes the weekly CDRs as well.

      Some tips for good "short term backuping"

      1. Use RAID-1 not RAID-0

      2. Use a good file system like reiserFS that's fairly immune to the side effects of being turned off in a hurry [ntfs and ext2 are not]

      3. Don't use the raid drive as a home or other frequently accessed directory. Do your backups from a well tested cronjob. The less you play with your raid drive the less likely you are to delete/mess with the files on it.

      4. Do backup the files from the raid to a medium like tape or CD/DVD on a regular basis.

      5. If one drive should fail, do a backup before you restore the raid. That way if you mess up restoring the raid you don't lose data [particularly beginners do this]

      Tom

      --
      Someday, I'll have a real sig.
  50. There's a lack of real information about RAID by mollog · · Score: 4, Informative

    I worked for years in development of RAID solutions for a major manufacturer. One of the problems with selling RAID solutions is the lack of understanding, or the prejudice and bias of the people who were supposed to be specifying and buying the hardware.
    The 'tutorial' of the parent article is talking in kindergarden terms, oversimplifications and obsolete term, and overlooking some of the issues with using RAID. It's a good example of the true lack of understanding about the subject. By now, there are so many types of solutions that the term RAID hardly applies. But, even 10 years ago companies like Compaq had innovative rudundant storage solutions that were enterprise ready.

    --
    Best regards.
  51. Re:HP 6402 by Anonymous Coward · · Score: 0

    My favourite failure was having the array accelerator memory fail (192mb)while I was working on a 2003 server. Unfortunately I was making changes to active directory at the time - and some of the drive writes happened to be in the array memory at the time.....ah bugger....

  52. Re:Give me RAID 5 AMEN by snuf23 · · Score: 1

    "If anyone will ever make RAID 5 a desktop standard its Apple. Its just too bad the G5 tower only has room for 2 drives."

    Now THAT is the most hilarious thing sentence I've read in a long time.

    --
    Sometimes my arms bend back.
  53. RAID virgin pops cherry... by krynsky · · Score: 1

    I just built my first RAID server. I've been the victim of 2 hard drive crashes where I lost data. I have a large media collection that includes family videos of over 30 DV tapes I've captured. Between all the media as well as other critical info I finally decided to dive in. Since basically all I needed was a file server that also offered redundancy I opted for RAID 5. I didn't need huge performance and in a RAID 5 configuration you only have to sacrifice 1 hard drive for redundancy as opposed sacrificing equal numbers for mirroring. This seemed like a good comrpomise to me. After making that decision I had to find a controller. I ended up going with a Promise Technologies SATA raid card. I mainly chose it due to the affordability and good reviews. Be careful of some cheaper cards that only offer RAID 5 in software. Another major consideration was the fact that I only had a standard PCI slot available in the Dell 400SC server I was using. That limited my choices quite a bit since most of the other contenders only worked on a PCI-X bus which requires a much pricier server motherboard. So my build went smooth and I am extremely happy with what I now have. Here's another good bit of advice that a friend passed on. Make sure to test your array after you are done by forcing a failure to understand how your hardware / software deals with it and you are aware of what to expect. I did this by starting to copy a 4GB file to the array and during the copy I unplugged the SATA cable from a drive. The software quickly alerted me of the failure. Shortly after I plugged the SATA cable back in and rebooted. The RAID bios also identified the failure. I then booted into Windows, launched the RAID utility and watched as it rebuilt the array. Everything went smooth and I now feel very confident about my server.

    1. Re:RAID virgin pops cherry... by jube_fl · · Score: 0

      For the future, almost all PCI-X cards will work in a PCI system.

    2. Re:RAID virgin pops cherry... by Diag · · Score: 3, Interesting

      If that RAID array is the only place you are storing all those home movies, I highly recommend you take a backup to some other kind of media. As other people have said, the RAID controller is a single point of failure. If you lose that, you lose the lot. And there's no guarantee that another controller will be able to rebuild it. Sad but true.

      --
      Serving Suggestion: Defrost
    3. Re:RAID virgin pops cherry... by typical · · Score: 1

      ...during the copy I unplugged the SATA cable from a drive. The software quickly alerted me of the failure. Shortly after I plugged the SATA cable back in and rebooted.

      Is SATA electrically safe to hot-plug?

      --
      Any program relying on (nontrivial) preemptive multithreading will be buggy.
    4. Re:RAID virgin pops cherry... by krynsky · · Score: 1

      Well to clarify, all of those home movies are still in their original format (DV) and the backup is more so that I don't have to re-capture and encode them. But the point is well taken.

    5. Re:RAID virgin pops cherry... by jpc · · Score: 1

      Is SATA electrically safe to hot-plug?

      Only with a proper sata backplane.

      Pulling the data cable for testing is I guess probably ok, wouldnt pull the power cable though...

  54. Absolute crap by Anonymous Coward · · Score: 0

    More ads than pages and the content sucks to boot. Why is this wad allowed to post this?

  55. RAID by TheScorpion420 · · Score: 1

    What I don't understand is why this made it to the frontpage, seriously how many people on slashdot don't know what a RAID is? I've known what RAID is since I was 13. What kind of crap passes for a submission nowaday anyways?

    --
    If you pay your taxes you support terrorism!
  56. RAID is just NOT reliable-Heat Death. by Anonymous Coward · · Score: 0

    I'm assuming all your "enterprise" RAID setups were IDE. RAID with a good card, and SCSI drives does excellent.

    The one issue I think every HD user should be more worried about is HD temperature. More drives fail due to excessive temperature than any other reason. RAID really isn't going to help if all the drives are running in an excessively hot environment.

    ---
    The "are you a script" word for today is selfish.

  57. At least one glaring inaccuracy... by wh0me · · Score: 1

    RAID 5 may be the most powerful RAID configuration for the typical user, with three (or five) disks required.

    Raid 5 can be configured with a minimum of three drives.

    Interested in RAID? Google, or check out: Storage Review's excellent information.

  58. It's really not that hard, people... by chriswaclawik · · Score: 1
    ... all you have to do is point the can at the beehive and spray.

    And for god's sake try not to get yoursefl stung.

    You don't really nead a whole article about using RAID.

    --
    A guy walks into a bar... well, I forgot the joke, but the punchline is that he's an alcoholic.
  59. Manual RAID by LennyDotCom · · Score: 1

    I have always used RAID M you have more then one HD in your system and you copy what ever is important to more then one drive. Since I use a Mac i also have the OS on other drives so if my OS gets toated I just boot from a different drive. Works for me for the last 10 years.

    --
    http://Lenny.com
  60. in other news by Anonymous Coward · · Score: 0

    a groundbreaking article describing how humans breathe air by breathing in and breathing out.

    www.this-is-not-news.com

  61. oh and i forgot to say... by Anonymous Coward · · Score: 0

    that zonk is a fucking retard. =]

  62. RAID Karma Drain Experiment 1138 by NotQuiteReal · · Score: 1

    You don't need backups if you don't do anything important. I should know. I am an official member of RAID - Redundant Army of Independent Derelicts. Of course, before PC (political correctness), back in my day, we just called ourselves HOBOS. (Huge Obnoxious BO Smellers). One day, I was wetting myself and talking to God, and he said, "Jed, get away from there, you don't need no RAID. Black gold, Texas tea.", Next thing you know, I'm a Lotto (TM) millionaire. So, I went to California and started a dot-com company. That didn't work out, but I don't miss it. Nothing has changed for me. To RAID or not to RAID, that is the question, I ask myself every time that wiener dog goes by... [Can you beleive they give 5 whole mod points to people who use public access terminals!]

    --
    This issue is a bit more complicated than you think.
  63. Re:Wow, a whole story I disagree with! Hmm.... by NotBorg · · Score: 1

    1. RAID implementations on most consumer-grade motherboards (EIDE RAID with Promise controllers on-board, and so on) are cheezy. I've tried using them for several years now, and I ran into lots of unexplainable "glitches" that never occured when I took RAID out of the equation. (EG. The RAID array would suddenly report a failed drive, yet when I'd pull the drive out and try reformatting/reusing it on another machine as a stand-alone C: drive, it would have no problems at all and S.M.A.R.T. reported it was fine too.)

    Yes. Promise controllers suck. I have one and occasionally it "forgets" how the drives were set up. Nothing is actually wrong with the drives. It is possible (on mine anyway) to re-enter the configuration and resume normal operation without data loss. The trick being is answering NO when asked if you want to quick initialize the RAID.

    I've had no problems with my other on-board RAID controller (made by VIA).

    --
    I want this account deleted.
  64. And all of your points have fallacies.... by postbigbang · · Score: 1

    If you buy junk, expect it to fail-- and that means anything sold by Dell to start with.

    1) RAID controllers-- real ones-- are highly reliable. Yes, a power supply surge can kill anything. Yes, backups are a good idea. But if you're in production, backups are stupid; high-availability clustering is important. Client/users that need RAID get a benefit.

    2) See the comment about using Dell products; they make nothing themselves, and therefore are the WalMart of computer quality-- driving OEM costs to the bone, and therefore, often quality, too.

    3) There's no research or logic that supports this. It's another computer urban myth.

    4) Once again, buying from the bottom of the barrel is an unwise posture for a serious professional. This isn't ToolTime.

    Backups are a good idea, They work when data can actually be restored in under a century. It's nice also to have mirrored external USB 2.X or FireWire/IEEE1394 drives, too. But a majority of people don't even do that. In production environments, you're destined to do it and for all the right reasons. But it's not a good excuse to re-install an OS. You do that only with Windows. A real OS can live a long healthy life.

    --
    ---- Teach Peace. It's Cheaper Than War.
    1. Re:And all of your points have fallacies.... by King_TJ · · Score: 1

      I think we're talking about two different situations. I thought the main point to the original story was how RAID is becoming a solution for "the masses".... EG. Home users and so-ho users.

      Dell is the #1 seller of computer systems, and therefore, there are a LOT of Poweredge servers out there - especially in smaller businesses. So simply discounting them all as "junk" is ridiculous. As you said yourself, they don't even really build anything themselves - so you're just as likely to run across things like their RAID controllers in other systems made by HP/Compaq, IBM, etc. - because they're all sourcing their parts from the same controller builders.

      As far as my hot-swap drive tray comment, I'm talking about what the average consumer can by off the shelf at a local computer store. Last time I checked, you simply couldn't get a good quality hot-swap tray for a clone PC case at CompUSA or any of the smaller computer dealers in town. Like anything, pay enough and search enough and you can find quality - but we're talking about the "typical" product you'd run across here. And they typically suck!

  65. Linux Software RAID and USB... by __aaclcg7560 · · Score: 1

    Has anyone tried using Linux Software RAID and a bunch of external USB drives? I'm wondering what the performance would be like for a light-duty file server. I want to go from a huge case to much smaller case, and that means getting the RAID drives out of the case.

  66. 1+0 or 0+1 by sgtsilver · · Score: 1

    I was looking through some raid articles and found this.
    http://www.ofb.net/~jheiss/raid10/
    It explains how 1+0 is better than 0+1 either way..ROACHES HATE RAID!!
    .
    .
    .
    ... sorry, had to throw in the corny bug joke.

  67. raid for desktop - not really worth it by vlad_petric · · Score: 2, Interesting
    RAID makes a lot of sense in a server scenario, where uptime is crucial, and where the cost of hardware is smaller than the business lost by downtime.

    As far as desktops are concerned - well, RAID and cheap just don't mix. For instance, if you just want reliability, RAID 1 is enogh (2 drives). If you want reliability + fast writes, you need RAID 1+0, which means 4 drives (RAID 5 only gives faster reads). Furthermore, a good controller is crucial (from my experience, these generally cost upwards of 100$).

    Finally, RAID does not subsume in any way a good backup system. I've seen cases where a damaged controller broke both harddrives in a RAID 1. However, for (most) desktop PCs, a good backup system does subsume RAID, since it's generally easy to just use a different computer, and get all the files from the backup.

    For me, the excellent piece of software backuppc running on a cheap box (~300$) has worked like a charm. This might not look cheaper than RAID, but considering that I'm using just one box to back up 10 other machines, it's pretty good.

    --

    The Raven

  68. My non-RAID backup solution.. by the_rajah · · Score: 3, Interesting

    No, it's not "Real-Time" but it suits our needs in our home office situation.

    I use "Smart Synch" software to incrementally copy the desired directories from the working computers to a "Backup server", an older Celeron machine on the network. Separate partitions are set up for each computer that is being backed up. At Midnight the incremental backups are made.

    Then at 2:00 a.m., Smart Synch running on the backup server makes another backup to a USB hard drive plugged into it. That USB HD is on a regular plug-in timer so that it only runs during the time of night when a backup to it is being done. The idea there is that the running time is limited and drive life is extended. Weekly, a backup DVD is burned and stored off site. Am I being anal? Maybe.

    --


    "Do the Right Thing. It will gratify some people and astound the rest." - Mark Twain
  69. It can be part of a backup technique... by sczimme · · Score: 4, Interesting


    Ultimately, what it comes down to is that mirroring merely makes the hardware more reliable, it is not a backup technique.

    It can be part of this nutritious breakfast^W^W backup technique:

    0) shut down the box

    1) swap a fresh/new/wiped drive for one of the mirrored drives

    2) rebuild the RAID

    3) store the just-pulled drive appropriately (e.g. off-site) along with a second identical RAID controller

    Now if the machine goes completely belly-up (as in a fire) the user can install the secondary RAID controller and the data-laden drive in a fresh machine, add another fresh/new/wiped drive, and rebuild the RAID in the new machine. This may not be terribly convenient nor perfect for everyone but it will be effective.

    Remember, kids: just because a particular technique doesn't perform a task all by itself (in this case RAID 1 != backup) that doesn't mean it can't be part of a larger picture.

    --
    I want to drag this out as long as possible. Bring me my protractor.
    1. Re:It can be part of a backup technique... by rhyno46 · · Score: 0

      It seems like a good idea, and an ideal solution. But, when you are flustered due to a disaster and you are carrying that disk back to your new server where you accidentally drop it 2 feet to the desk...now you are fusked. All of the sudden your super fast backup is trash. HDDs are just too touchy for me.

      I just can't seem to find a good alternative to the reliability of multiple tape backups.

  70. Is this for availability or currency? by gelfling · · Score: 1

    I mean if you just want fault tolerance then use straight up mirroring. If it's for currency then perhaps you need a journalling file system with checkpoint rollback/rollforward.

  71. Why fakeraid really sucks by Lost+Found · · Score: 2, Insightful

    My father has trusted his data (against my advice) to fakeraid chipsets on his various motherboards twice. He just got done *losing* all of his data for the second time.

    Best we can tell, he had one drive go without his RAID controller warning him; then had a second drive go, killing the array. He spent weeks with a dead PC playing with all kinds of special Windows bootloaders and disk recovery tools trying to get his files back.

    Fakeraid sucks because it's just a line item on the sale of a modern motherboard. The inclusion of the "RAID" functionality is borderline fradulent. REAL RAID controllers, of course, have a coprocessor and often battery backup and leave all of the storage details to themselves rather than some fly-by-night driver in the operating system.

    It's no surprise that they come with virtually nothing in terms of recovery software.

    The happy median, I've discovered, is Linux md. md supports many RAID levels, and according to some benchmarks will certainly outrun fakeraid in performance (which doesn't particularly surprise me). The administration tools let you simulate drive failure, monitor array health, create degraded arrays, and the documentation tells you what to do when something goes wrong.

  72. Re:Wow, a whole story I disagree with! Hmm.... by Diag · · Score: 1

    I also disagree with the story. People always confuse RAID with backups.

    RAID provides *redundancy and availability*. If a disk fails, you can get your system up and running again almost instantly - obviously very important in the business world.

    Backups provide *data protection*, so that if your data is lost or corrupted for any reason, you can retrieve it from a safe copy.

    The only purpose I can see in the home for RAID would be RAID-0 (striping) for performance improvements. But the only home application I can think of that would utilise that improved sequential throughput would be video capturing/conversion. Most home applications are random I/O intensive, which doesn't get much improvement from striped disks.

    For backups at home, most people should be able to get by with occasionally burning a CD or DVD of their really important documents and records and storing that at another location. If people *really* feel they need a backup of those "linux distributions and home movies" that are occupying hundreds of GB on their hard drives, external USB hard drives are probably the best option at the moment.

    --
    Serving Suggestion: Defrost
  73. Raid 10 is cheaper than RAID 5 by erice · · Score: 1

    Raid 10 needs 4 drives but a much less expensive controler. Adequate performance for Raid5 means a hardware XOR engine. Such a controler starts at about $150. A 4th 200GB drive is available for less than $100.

    RAID 10 is both faster and more reliable than RAID 5. (Though the chance on losing data for either is mighty low)

    RAID5 really only makes sense for arrays built out of the largest disks available.

  74. Why wait? Everybody waits.. by Man+in+Spandex · · Score: 2, Insightful

    Losing data once due to hard drive failure may be all that is required to convince anyone that RAID is right for them, but why wait until that happens.'"

    Isn't it human nature (or at least that's what it seems) to wait until something "bad" happens?

    That goes for obese people, smokers and yes even computer geeks.

    Why eat all the fat? I'll just burn em all!
    - Wait til your 40-50 and check that cholesterol strike...

    So many people smoke and get away with it so I will to right?
    - Yeah wait til you get some health problem that will make you say "OH NOES!"

    Why I need firefox? ActiveX hasn't screwed me.
    - A week later "omfg whats all this junk, I want Firefox!"

    Why Do I need RAID or even a burner? I got 3 hard drives that contains all my data!
    - 8 months later, 1 hdd crashes "AH F*K ALL MY Pr0n!" and then he thinks of having a simple RAID 1 setup...

    We always wait because we are lazy and cheap.

  75. First FireFox .0.6 Bug? by RKBA · · Score: 0, Offtopic

    I wonder if I'm the only person to file a FireFox bug report that the "Read More" link to this article on the front page kills the new version of FireFox? For some reason, this URL aborts FireFox 1.0.6:

    http://hardware.slashdot.org/article.pl?sid=05/07/ 22/2159230&tid=198&tid=1&tid=218

    Oddly enough, the link works fine if you copy/paste it into FireFox's address box, but not if it's clicked on from the main SlashDot page. Can anyone confirm this?

  76. Which RAID controller is good? by swatoa · · Score: 1

    I had high hopes for nvidia's nforce4 onboard RAID, but after someone said it sucks, I don't know what to look for. Are those $100 Promise cards good game?

    1. Re:Which RAID controller is good? by vlad_petric · · Score: 1

      Hmmm, I don't know the promise cards. I've been using a HighPoint PCI-X RocketRAID 1820A (~210$), and it's been working pretty well. Here's one review that recomends it for desktop: tweakers. I'd do some more research before purchasing a RAID solution, though. What you basically want is a "smart" RAID card - one that does all the work by itself, without asking the main processor to compute stuff (like RAID 5 XORs).

      --

      The Raven

  77. No competent editors on duty? Don't post articles! by Stankatz · · Score: 1

    This article focuses on the the basics of RAID...

    The the the the editors must be be be be blind. How long ago did you give up giving even the most cursory inspection to the summaries? I know you all think it's only the headlines that matter, but some of us (i.e., all of us) readers wish you would do a little more, you know, editing.

  78. Raid 10 is cheaper than RAID 5-On eBay! by Anonymous Coward · · Score: 0

    "Adequate performance for Raid5 means a hardware XOR engine. Such a controler starts at about $150."

    I must be the only one who knows how to shop for computer gear.

  79. 'RAID' basics: You don't need it. by Anonymous+Freak · · Score: 2, Insightful

    1. The average user does not need RAID.

    2. The enthusiast does not need RAID.

    3. RAID is not a replacement for backing up.

    RAID is only good for two things: To protect against a hard drive failure when an uptime as close as possible to 100% is required; and to increase performance in the form of data throughput.

    Alright, let's cover them one at a time. The average user (anyone who just buys a Dell to surf the web,) doesn't need RAID. Hard drive failures are fairly uncommon. And the to the 'average' user, 100% uptime isn't anywhere near a necessity. Also, data throughput isn't that important, either. (Besides, for most people, buying a faster hard drive is both technically better and more cost effective.)

    Most enthusiasts (gamers, hardware tweakers, modders, etc,) don't need RAID. Again, hard drive failures are fairly uncommon, and your average enthusiast doesn't exactly store cures to cancer or rocket science on their drives. (No, the old MS Space Simulator doesn't count as rocket science.) Besides, most enthusiasts use RAID-0, which isn't really redundant. Which brings us to point B. The only thing RAID-0 measurably improves is sequential STR speed (Spindle-to-RAM, the speed of data going from the platter to the drive's internal cache.) And there are very few enthusiast tasks that do better from a raw higher throughput. (Even capturing video doesn't matter, as the slowest hard drive today can easily keep up with uncompressed HD!) Oooh, so your Doom3 level loads 1/2 a second faster. In single-player mode, it doesn't matter at all, and in multiplayer, the server waits for everyone to load anyway. RAID doesn't help for random seeks. A faster spindle helps for that. A better drive caching algorithm and a larger drive cache helps for that. If you have a 40GB, 5400RPM, 2MB cache drive, you'll see a tremendous improvement by going to a Raptor, or to a 15k RPM SCSI drive. You will see very negligible benefit by getting a second 40GB, 5400RPM, 2MB cache drive and putting them in a RAID-0.

    Finally, the assertion that RAID prevents data loss. The only data loss it prevents is loss due to a failed hard drive. It doesn't protect against user error, viruses, or physical damage that destroys the whole computer. If your data is truly vital, you need to be backing up, even if you do use a RAID. Yes, businesses with vital information who need (as close as possible to) 100% uptime need RAID. That's it. Even then they still need backups.

    --
    Another non-functioning site was "uncertainty.microsoft.com."
    The purpose of that site was not known.
  80. Personally... by steelfood · · Score: 1

    ...I like this guide better.

    Yes, it's a little old. But, I found it to be very comprehensive, especially for its day (remember, RAID 0 and RAID 1 were only beginning to show its face in the consumer market when the guide was written). It's actually a mirror of another page, but since I stumbled upon this one first, I don't remember where the original is or what it might have been called, or, for that matter, which pages/paragraphs were in the original and which ones were not.

    --
    "If a nation expects to be ignorant and free in a state of civilization, it expects what never was and never will be."
  81. the article is full of errors by erc · · Score: 2, Informative

    One glaring error:

    RAID can be run on any modern operating system provided that the appropriate drivers are available from the RAID controller's manufacturer. A computer with the operating system and all of the software already installed on one drive can be easily be cloned to another single drive by using software like Norton Ghost. But it is not as easy when going to RAID, as a user who wants to have their existing system with a single bootable hard drive upgraded to RAID must start from the beginning. This implies that the operating system and all software needs to be re-installed from scratch, and all key data must be backed up to be restored on the new RAID array.

    Again, wrong, wrong, wrong. There are hardware RAID 1 controllers that require no drivers and you don't have to do squat - just power down the server, install the RAID 1 on your IDE interface, plug in the new drive, hit the power, and away you go. The controller is smart enough to automatically sync up the two drives in the background.

    --
    -- Ed Carp, N7EKG erc@pobox.com PGP KeyID: 0x0BD32C9B What I'm up to: http://intuitives.mine.nu
  82. Yes, they do... by RKBA · · Score: 1
    Do people wear bulletproof vests?

    In my neighborhood, yes they do! ;-)

  83. Slashdot as a source of info has peaked by linzeal · · Score: 1
    Yeah that is why I don't come to Slashdot anymore cept by RSS link to article as there are dozens of better sources for news and information out there now. The only thing slashdot has going for it is mass of people. If you want real science news that is not days old rehash with ads try Eureka Alert for real science and The Inquirer for computer industry news.

    The other problem is that most people who work at Slashdot have very limited skills outside of the core competencies it takes to run a web forum so we tend to get uninformed blather and the publishing of stories that even simpletons with science degree would not publish.

  84. Some notes about the aritcle by Torg · · Score: 1

    In the article a few things are omitted that would be important to note if you were building your own array.

    1) The article notes raid 0+1. Raid 0+1 is VERY dangerous. You want 1+0 not 0+1. The main reason is loss of any disk from a 0+1 will immediately degrade your array. Also keep in mind vendors call this 1/0, 1+0, and 10. Also keep in mind vendors LIE about 1+0 vs 0+1

    2) The alternative to 1+0 is raid 5. Raid 5 had some significant drawbacks to its use. These are read-modify-write and the calculation of checksums. In fact, unless you get higher end raid controllers, it is incredibly slow. Software raid 5 is even worse. So be careful here too.

    Since none of this is new groundbreaking material by all means go look up whatever storage solution you decide on implementing. Learn how each solution works and your normal utilization.

  85. backup concerns by sl4shd0rk · · Score: 1

    Raid is big storage for cheap - especially SATA Raid. A hardware card can be had for under 300 bucks for a 6 unit LSI. For a neat experiment, try Linux software raid booted from a Knoppix 3.9 CD. On a regular IDE subsystem, you can have your raid 5 which is independant of the os layer. The big concern is backing up that big 3 drive, 320G raid device once you fill it up. Especially if your a corp and you want to level 0 all the invested time/money that is laying on that device. It can get quite pricey, and Amanda will not span a level 0 to multiple tapes.

    --
    Join the Slashcott! Feb 10 thru Feb 17!
  86. Nice, but not what I am talking about by meehawl · · Score: 1

    a decent case is a much better option...

    That's a pretty nice case, but firstly it's not available yet, secondly it's a PC case, awaiting a motherboard, power supply, RAM, and so on, and thirdly, a cheap firewire 4-bay JOB enclosure costs around $70, all in, power and fans, and you're good to go.

    The annual power consumption difference between running an enclosure or rack of disks and a whole intelligent CPU storage node is substantial: on the order of ~$100 annually. Not to mention it's also more of a PITA to support, with many more parts liable to failure or corruption. When I want to complexify my life by setting up intelligent nodes, rather than just slaving drive boxes, I usually go with a good Antec fileserver case.

    --

    Da Blog
  87. Re:Excellent RAID reference: not by Jamesday · · Score: 2, Informative

    Consider:

    RAID 10 disadvantage: "All drives must move in parallel to proper track lowering sustained performance". In fact each drive can seek independently for reads and only pairs must seek together for writes.

    RAID 1 advantage: "Transfer rate per block is equal to that of a single disk"
    RAID 5 disadvantage: "Individual block data transfer rate same as single disk"

    Would be nice if it was consistent about whether that's good or bad.

    RAID5: "Highest Read data transaction rate" except for RAID 10, of course, where you've less chance of being bottlenecked because there are two sources for each stripe.

    RAID5: "Medium Write data transaction rate", only the lowest of all except 50, because of the parity calculating and writing to a second drive.

  88. RAID 1 for read speed!!!! by sygin · · Score: 1

    I have installed software RAID 1 (mirroring) on a system with two SATA drives 2.6 kernel. It was easy with the new debian installer in Sarge.

    The speed increase of the system was wonderful. Boot times, responce times etc. Write times remain the same but read times are halved.

    I have since upgraded all our Linux servers to PATA Raid1 - Requires no raid controller. Query times have dropped. All for the price of a few 40GB IDE drives.

    I now want RAID 1 for my home machine.
    For the boot times etc not just data redundancy.

    --
    Don't make your problems my problems!
  89. Cheap RAID - No controller hardware required. by sygin · · Score: 1

    I have installed software RAID 1 on my Linux systems
    and they boot faster. No hardwaer required, uses linux kernel software RAID - much faster than hardware raid on a dual processor system.

    Works for IDE and SATA drives. Easy to do with new Debian Sarge installer.

    --
    Don't make your problems my problems!
  90. Works well on insects, too. by i41Overlord · · Score: 1

    It's a disk redundancy technology and an insect killer.

  91. TLDP has free docs already by Anonymous Coward · · Score: 0

    The Linux Documentation Project has at least 3 free documents not infested with advertisement. Can I submit these too, one at a time??

    That would at least given valuable comments and feedback to improve them.

  92. Hardware v. software RAID by Rich+Klein · · Score: 1

    No mention of hardware versus software RAID? I know the article is geared towards noobs, but I think they should've mentioned the pros and cons of hardware v. software.

    --
    -Rich
  93. Re:Wow, a whole story I disagree with! Hmm.... by mink · · Score: 1

    "1. RAID implementations on most consumer-grade motherboards (EIDE RAID with Promise controllers on-board, and so on) are cheezy. I've tried using them for several years now, and I ran into lots of unexplainable "glitches" that never occured when I took RAID out of the equation. (EG. The RAID array would suddenly report a failed drive, yet when I'd pull the drive out and try reformatting/reusing it on another machine as a stand-alone C: drive, it would have no problems at all and S.M.A.R.T. reported it was fine too.)"

    I bought a KT7A-RAID when it came out (been something like 3-4 yers now I think) and it uses a Highpoint chip. I have never encountered any issues with the RAID config, I have used it for RAID0, RAID 1, and both.

    --
    Well I've wrestled with reality for thirty five years doctor, and I'm happy to say I finally won out over it.