Slashdot Mirror
IDE RAID Examined
Bender writes "The Tech Report has an interesting article comparing IDE RAID controllers from four of the top manufacturers. The article serves as more than just a straight product comparison, because the author has included tests for different RAID levels and different numbers of drives, plus a comprehensive series of benchmarks intended to isolate the performance quirks of each RAID controller card at each RAID level. The results raise questions about whether IDE RAID can really take the place of a more expensive SCSI storage subsystem in workstation or small-scale server environments. Worthwhile reading for the curious sysadmin." I personally would love to hear any ide-raid stories that slashdotters might have.
Don't tell me the hard drive manufacturers are in that much trouble. I hope Willie Nelson will be there.
IDE can only handle one or two hard drives per channel, which makes the cabling a real nasty hassle as opposed to SCSI-based RAID.
Even those so-called rounded cables can clutter the hell out of a tower case if you have a 4-channel RAID controller.
In my case it's the Adaptec 2400A four-channel, with four 120GB Western Digital hard drives, RAID 1+0.
Whats the point in having SCSI-Raid in most workstations these days? I mean, ram is so cheap now you can throw in a couple gigs for much less then the price diffrence between SCSI RAID and IDE raid.
I mean, I know the hest drives are SCSI flavor, but it seems like there's so many other things you could spend money on first that would get you way better performance, like getting a Dual Athlon CPU or something.
autopr0n is like, down and stuff.
That's so true. There's a big difference :-).
Sex - Find It
At the company I work for, IDE RAID has become somewhat standard because we're basically cheap... At least it's standard on the servers that are fast enough to support it. The rest use dd to copy partitions between backup drives. My boss calls it "RAID point five" We lovingly refer to it as the ghetto network.
I ran an IDE RAID, one of the first, a few years ago. It was a 3ware RAID-1 controller. I thought it would be useful because I had gotten sick of losing data on a drive failure. I didn't have the money (or patience :) for a good backup solution and Linux RAID hadn't matured.
... never have I seen 2 drives go down simultaneously. Nor have I seen a controller malfunction in a way that damaged the drives (though I've heard of it from other people).
Everything was fine for awhile. After a few months I lost a drive, replaced a drive and it remirrored fine. Same thing happened a year or so later.
Then one day my controller fried. Nothing else in the system went down, but some kind of surge hit the 2 drives from the RAID controller. The controller still worked but neither drive was accessible, either as RAID drives or as single drives. Tried numerous tricks, eventually gave up.
I've run SCSI RAID in boxes I admin at work
All in all, I decided it wasn't worth it. I am currently doing Linux mirroring in combination with journaling filesystems on one box, and Windows mirroring on another.
It is more productive to voice thoughtful opinions (reply) than to judge (moderate) others.
---- El diablo esta en mis pantalones! Mire, mire!
offtopic? Who gives a shit about apple?
I've got a RAID-5 machine made with 5200 RPM WD 120 GB drives. Works great. It's a light server, and I built the thing for under 700 bucks, dual procs and all.
I didn't use a RAID card, just a couple of IDE cards. And it was amazingly simple to set up.
Linux - Because Mommy taught me to Share.
I personally would love to hear any ide-raid stories that slashdotters might have.
Once upon a time, in an array far, far away, there lived a young princess who was worried about the integrity of her data...
For us small-time sysadmins, RELIABILITY is the concern. Can I yank a drive, pop in a new, unformatted one out of the box, and expect to be able to replace the other drive in a similar fashion the next day, without losing a single file?
Don't give a shit about Q3 times. Do care about a RAID Suse Linux server. Remember, I'm not some millionaire Slashdot commer with greasy palms and a red cock, I have chilluns to feed.
It's spelled r-e-l-i-a-b-i-l-i-t-y, advertisers.
My favourite quote from the article : As an added bonus, the lights sometimes flash in a side-to-side in a pattern reminiscent of Knight Rider's KITT.
I'm working on a project where we use an IDE Raid card for data backup on a computer/device we ship to customers. We use a simple mirror, basically just insurance against a hard drive crash. It's allowed us to keep our costs much lower than they would have been with SCSI. We have only had to recover one time, but it was very easy.
Brevity is the soul of wit
-- Polonius
The only downside I can see with raid is one main thing for me. After I went raid I couldn't go back to standard ide. =) I know I "could" but it would be a pain in the a$$. Lets say I start to build a backup computer and decide I want to take out one of my 80 gigs and pop it in the new one. Sine my two 80's are currently in a raid I think this would have some seriously messy side effects. Thats my only bitch about raid. Other wise its pretty damn cool. Also my floppy drive is broke so I cannot boot windows xp to install on one of my raid drives. Sucks. Stupid xp -).
.[[erax0r]].
Unfortunately, I don't think they tested the built-in RAID capabilities of the Xserver. Or, Apple add-on cards in general. So it very well may be that IDE on Macintoshes is better than SCSI anywhere. This article has no comment on that.
How the Cheapest card (the Highpoint RocketRaid 133) seems to be the best in the performance chart.
Maybe it's just me, but why woould I go back to SCSI when IDE almost outperforms for about a third of the cost.
Just because I'm a geek doesn't mean I'm made of money.
I've had a 3ware Escalade 6200 (ATA/66, two IDE interfaces) running in an RAID-1 (mirroring) for about two years now, and I haven't had a single problem with it running under Debian Linux, kernel 2.4 series. It's nice to have the piece of mind RAID gives you. I also have a somewhat beefier server with a more recent 3ware Escalade 7400 (ATA/100, four interfaces) and again, no problems and great performance. I would definately recommend 3ware, but I have not tried too many others (haven't needed to.) There future offerings of their Escalade line combined with Serial ATA should be interesting to watch once SATA drives become available.
Pete
Well...apparently I do...uh, you do.
Or, maybe they are just buying SCSI cards...
I'd be happy if I could find a decent external IDE RAID enclosure at a good price. So far, the only ones I've seen cost waaaaaaaaay too much money. Is there anything similar to a Sun 711 Multipack for IDE? (Hopefully something I can buy on the cheap through Ebay?)
You would think that after 130 graphs comparing the controllers he could come up with a stronger conclusion than "I cant really decide which one is the best"
"The defense of freedom requires the advance of freedom" - George W Bush
I have two 40 gig WD HDs that were originally purchased for RAID, but I've not had a wild hair enough to use them. So I've got an E: and an F: drive and occasionally back up the E to the F. So, in summary, I have RAID capabillities, but choose not to use it, because 7200 RPM on one drive is fast enough for me.
I know, what a shitty story. I feel like the videocam/train guy.
A friend of mine set up a raid0 (striped array) using the built-in raid-controller in his motherboard. Later, this motherboard had to be changed. To our great surprise, the raid information was only stored in the motherboard and thus permanently lost. This could be a good thing to know ... Make sure the data is not lost if the controller fails.
Personnally, I run several software RAID arrays under Linux and it works very well. It's easy to manage and gives me decent performance on my rather old machine.
I feel very confident in mirroring system/boot partitions on my linux machines =)
I use scsi (raids) only at our production servers (sysadmin for a medium sized company in the UK). We run a database intensive website serving up to a 100k ppl per day.
:)
While these drives may be expensive (15k rpm / 18 gb / 4-6 per machine) they definately dust the arena with any IDE-raid I've ever seen. But then again I've used only one IDE raid (something cheapo) for one other productionserver, and other than that just talked to friends who run it.
What really would be interesting to see would be the latest and greatest from the IDE/raid-scene pitted against dito in the SCSI/raid-scene. All along with prices for controllers and disks ofcourse, to show how much more you have to cough up to get the scsi solution.
Hey why don't y'all paypal me 10 pounds each and I'll put a test like this online in a few days!
The jist of the story is that I lost almost 240 gigs worth of files. Yeah yeah, I did RAID 0, when one day one of my HD's decides to click incessantly. There was nothing I could do. I sat there tearing up as the drive was clicking because I knew I wasn't going to be able to save the drive.
I habitually backup files, but mostly the dynamic stuff like email, documents and ICQ databases. Most people don't backup 240 gigs worth of data, let alone a gig. I tried to restore what I could, but some stuff was lost forever. It was like a fire going through all my files. I guess that's a painful lesson learned.
RAID system bootup times are also longer, by a factor of 2x. But that's another story.
2 Western Digital drives in RAID 0. Fast as hell. Beats 15k SCSI drives.
I think that most of the innovative work is being focused on IDE, so we get uber fast drives like those (promised) for SATA. Meanwhile, the quality work is focused on SCSI, since they are a slightly more niche market and cost matters less. So, either get a really big, fast, innovative drive that might or might not die in a few months, or get a really expensive SCSI drive that will probably last years, but was obsolete (capacity wise) months before it was even purchased.
The way I see it, if one of my IDE drives goes, aside from being SOL, I'll just buy a new one. Hell, they're cheap enough...
I guess it's time to get two more drives and set it up for 0+1... just jinxed myself maybe?
Seriously, how can 50 people possibly read an in depth article on IDE RAID drives, and make any kind of intelligent comment on them, faster than I can type a silly little post like this.
www.arcoide.com - OS independent (including DOS!)
I have two linux file servers in my basement, each with a 4-port 3ware RAID5 array on it. One has 4x120GB drives and one with 4x80GB drive. Great storage for network-available (via Samba) MP3s and movies and the RAID cards were only $99.
It just works well.
Yes, RAID5 is a little slow on writes, but generally that's not the way my data flows. For a bigass box with REDUNDANT arrays, it's a very nice and INEXPENSIVE way to make that happen.
It has my hearty recommendation.
My experiences with IDE RAID have been pretty darn good. Benchmarking my Desktar 60GXP drives in Windows 2000 last year showed that I was getting read speeds in striping mode (between two drives) at faster rates than the fastest seagate Cheetah SCSI drives. Times have probably changed now though.
I started with a KT7A-RAID mobo. The important thing is that you get the cluster sizes just right for your particular partition. I used Norton Ghost to image my drive and try all sorts of different variables. In the end I had very satisfying results. Since I switched to Linux, I stopped using RAID-0 (yes, it is supported with this device!). I found that ReiserFS and the multi-drive Linux filesystem on these drives seemed to be just about as fast without having to hassle with soft-RAID controllers. It is probably due to my system RAM though. I couldn't seem to get Windows 2000 to make the most of 1024 MB without using that swapfile. Linux seems to avoid the swap altogether and uses static RAM instead. It is very nice having the extra IDE channels though. Without them, I probably wouldn't have 4 HDs hooked up right now.
It just means that the first 24 people didn't thorougly read the article.
In Soviet Russia, joke rides you!
I can't remember how I got by without IDE RAID ...
In fact I love IDE RAID so much I reccommend it to everyone I see on the streets ...
I even bought one for everyone in my family, just in time for the hollidays ...
Thank you IDE RAID, THANK YOU!
Ignore the "p2p is theft" trolls, they're just uninformed
I bought that about a while ago when the maxtor 160GB 5400RPM drives started to ship.
:) Reminds me when I plugued my first gigabyte drive in my amiga and saw big numbers :)
:).
I had to build a datacenter and storage price was the main issue. I had to have something cheap, yet hold a LOAD of data. Problem is personally I hate maxtor drives, I always found the more or less reliable (but drive experiences varies from a person to another so..). Anyways at that time maxtor were the only one offering 160GB drives, at a decent price/meg, and although 5400RPM is quite slow for access time, the main issue was cost so I could take a hit on access speed as long as "streaming" speed was fast enough.
the Adaptec 2400A card was the best at the time, simple, cheap efficient, it had 3 bad sides for my application, no 48Bits LBA support (130GB+), no 64bits PCI version (I was using a K7 thunder, and that chipset will slow down the pci bus to the slowest card connected to to bus, and since I wanted all available bandwidth to be thrown to the 64bits gigabit card, I couldn't accept using 32bits), and finally, no more than 4 drives. I wanted to break the terrabyte limit, so let's say I would have used 2 of those cards, it wouldn't have been price-performance-wise since the 2 would have shared the bus and I would have lost 2 drives for raid-5 instead of one with a 8 drive setup. but the performance of the Adaptec 2400A was the best. Still looks like the best overall today, yet I dunno if they are supporting 48bits LBA?
Anyways the 3ware 7850 was an excellent choice. Although their tech support is more or less good (like most tech supports) especially for real bugs and not just standard drivers reinstallation issues, the response time and sales people were very nice and professionnal. I got surprising results from the array, where I thought it would run like molasse, I was getting over 50MB/sec sustained non-sequential reading if I recall correctly. And the tools are very good, rebuild time is about 3-4 hours with 8x160GB @ 400GB filled on the drives, there are email alert tools and web interface to the host machine to check diagnostics. Overall it's a nice system and I'm sure the 7500 series are even better.
Oh and on a "funny" note, windows shows 1.1TB available in the explorer window, not 1134GB
As for the maxtor drives, I didn't take any chances, I ordered 10 to get 2 spares, 2 blew off in less than a month, but didn't have any problems since then, I guess if you can afford the time, doing a 1 month burn-in test with non critical data isn't overkill. usually they SHOULD blow up one by one so you could rebuild the array
--- Metamoderating abusive downgraders since my 300th post.
Thats bullshit. Post some links to benches that back that up.
Two 80GB WD special edition drives in RAID 0 (7200RPM, 8mb cache) rarely burst over 90MB/s. They usually have a sustained transfer of ~50-65MB/s.
Additionally, your seek time is going to suck. I gaurantee its not going to be under 11ms. You cpu utilization during transfer will prolly be around 4% in the asolute best case senario and 11% on average. This is becuase, no matter what you think, all raid cards under ~140$ do the calculations for the transfers in software, not hardware. All you have is a controller card with special drivers. You wont come even close to beating the overall performance of a scsi 160 drive, or SCSI 160 RAID 0 setup.
HP has developped a pretty cool type of RAID. An automatic RAID-level that automatically organizes your disks for best performance while maintaining security.
When a friend explained it to me, it sounded like a mixture of raid 5 and 0+1. For example, if you replace a disk with a larger one, the extra capacity will be used to duplicate some other part of the array.
White papers here
Despite the claims of the article I have yet to see any 3Ware or HighPoint IDE RAID drivers. The simple reason being that they do NOT want to release their IDE RAID source code as this is the IP their entire company is based on. The best you'll get is a driver with a precompiled binary of the IDE RAID engine and an open source wrapper. At least Promise is honest about this fact to the Linux community while the others just lie.
We've run several big RAID-5 setups on 3ware cards. When I say big I mean 1TB+ on each card. To do this we've used the 100GB+ drives available (120GB - 160GB) The biggest problem has been drive failures. Out of the 40 drives I think we've lost 6 in less than 1 year. In only 1 case have 2 drives gone bad at once (RAID-5, we're covered if 1 drive fails), but lost around 1TB of data. Luckily the data could be reproduced but took two weeks to regenerate.
It's WAY too easy to build massive arrays using these devices. How the hell are you supposed to back them up? You almost have to have 2, one live array and 1 hot spare array. If you think you're going to put 1TB on tape, forget about it. If you have the cash to buy tape technology with that capacity and the speed to be worthwhile, you should be buying SCSI disks and a SCSI RAID controller.
Using IDE Raid is like using a winmodem. Unlike with modems, where everyone has one, RAID has a basic educational entry point. I seriously doubt IDE Raid will ever overtake SCSI in any area where knowledgeable people are doing the administration.
Well - my computer has two 7200 RPM 80GB drives in a RAID stripe 0 array. Works beautifully. In Windows that is. I've been trying to install any distro of Linux on it for the last couple of weeks. All but one of the distros that I've tried couldn't find any RAID controllers. Mandrake found both hard drives - treating them as two - and not one drive. I already have set aside 20GB free using partition magic. So, anyone know which distros support this controller? It's part of the Gigabyte GA-7VRXP mobo. Thanks!
There are RAID systems that make use of cheap IDE disks but still have a SCSI connection to computer. Until consumers wake up and start demanding that IDE and SCSI disks cost almost the same, I think this is an ideal solution. It works well and it's both fast and cheap.
Addenda: ... and make a post in 8 minutes? ... less than, say, 30 minutes old...
/. posts.
Sorry about the errors, but this is my point. I'm just typing and I can't keep up. I can't even find my own posts without doing a text search now.
I guess I now know why there are so many typos and misspellings in
I have about 5 TBs of RAID5 storage online at various customer sites. They are all using Linux software RAID and Promise ATA66/100/133 controllers. Even when using two drives per IDE channel, we still see very good performance. An RAID5 system with eight 120-GB 5400-RPM Maxtor drives gives about 55 MB/sec write and 80 MB/sec read performance under Bonnie. Those eight drives were on two Promise ATA100 controllers. Cabling is fairly easy if you use 24" UltraATA cables. And it will get much easier with Serial ATA.
One customer ordered a system from a vendor who insisted on installing an ATA raid card, and it was a remarkable disappointment. Linux was able to indentify the array as a SCSI device and mount it. Then, for some reason, the customer rebooted his system. During the BIOS detection, the raid card started doing parity reconstruction and ran for over 24 hours before finally allowing the system to boot! For comparison, the same sized array would resync in the background under Linux in about 3 hours.
Also, the reconstruction tools built into the raid cards are pretty limited. If you have a problem with a Linux software RAID array, at least you can use the normal low level tools to access the drives and try to diagnose the problems. Just MO.
Has anyone used these Addtronics IDE -> SCSI ?
... Governments are instituted among Men, deriving their just Powers from the Consent of the Governed...
I have a highpoint ide raid controller built into my motherboard, and I must say it sucks. If the mirror set breaks, you have to put a new drive in, go into the bios, and remirror the drive.. this process took over an hour with 120gb hd. This sucks, because I want this to run live on our free-bsd mail server, and I can't afford to be down for an hour while the thing rebuilds. Then I tried setting up a 3 drive raid-5 system, and the damn thing wouldn't even boot when I pulled a drive out. This made me angry.
Here's a mod I posted before that converts a cheap Promise ATA-100/133 or ATA-66 controller into a RAID unit. http://www.tweakhardware.com/guide/raid100/ The last time I checked, Maxtor was selling the Promise unit as their own brand as well. This means that it's in wide distribution.
I'm a Dead/Phish/etc live taper, so I deal with fairly large (500MB-1.2GB) WAV files. I use an audio-enabled DDS drive to read in the data, Soundforge 5 to edit the files, and a CD burner to output the finished disks. My configuration is:
ABIT BX133-RAID motherboard with the onboard HPT370 RAID controller.
Pentium III 550
512MB RAM
Two 60GB 7200RPM ATA/133 Maxtor drives (running at ATA/100 due to the HPT370 limitation)-- one on each of the two supplied IDE channels, as a dedicated data disk.
One 20GB 7200RPM ATA/133 Maxtor drive running on a PCI ATA/133 non-raid ATA adaptor card (as the Windows system disk)
When I open a file in Soundforge 5, the software reads the entire file in at once. This can take several minutes for a large file. I recently switched from a single 60GB 7200RPM Maxtor drives to a pair of the same drives running in RAID 0 mode. (striping) The speed increase was dramatic. It didn't quite reach the theoretical maximum of a 50% speed increase, but but it did reduce the load time by about 45%.
Of course, I'm doing sequential reads -- the poster child application for striping. If you have an application such as digital audio processing that requires sequential disk access to large files, then using an IDE RAID controller to stripe your data can make a big difference.
RAID 5 can be a pretty poor performer, even with a dedicated RAID card with processor and cache memory.
(writes in particular)
I can't imagine doing software RAID 5, as the overhead is quite high.
I don't think SCSI for workstations is a good idea. I used to have a Compaq SP700 with 10K RPM SCSI drives and only after a month of use, I threw out the SCSI drives and got a Seagate Barracuda IV instead. What a difference! I would never go for a SCSI drive in a workstation.
In a server, I wouldn't ever go for IDE.
Promise controllers have a quirky setup display. About two years ago they said they would fix it, but haven't done that.
Anyone have comments about the others?
In any case, we use software RAID-1 so that the system can survive a drive crash. We started using RAID-1 on SCSI with the AIX Logical Volume Manager, and began using Linux RAID-1 on IDE when the Promise PCI controllers were supported in RH72.
We have lots of AIX and Linux systems, and have had a dozen drive crashes over the years.
and the author can't recommend a card? So much for skipping to the end like I do with most of these comparisons.
I was surprised that the author could recommend the overpriced Adaptec card for anything, At that price I expected better performance... no wait, I didn't.
Awesome!
I'm curious, so if anyone can help it would be appreciated.
I read the blurb about RAID 0 and striping, which intrigued me, but those damn graphs gave me a headache,
so performace wise, over a wide range of uses, (from development to gaming to serving)
would i get better performance with a dual drive IDE RAID 0 or just a single IDE drive hooked to a reg. IDE slot?
Having made the investment, I'll be wringing every last drop of sweat out of my homebuilt Linux/SCSI-160 network attached storage array thank you very much! I'm hoping that by the time that is on its last legs I'll be able to drop in a SerialATA RAID controller and a whole bunch of cheap drives to build the multi-terabyte storage array everyone will inevitably want by then.
UNIX? They're not even circumcised! Savages!
Holy cow. Sistina LVM (Logical Volume Manager) rocks. It is a partition system/file system of the future that really makes RAID sort of unnecessary. It is true that it is done by the host OS, but when integrated right it does not matter.
Documentation for LVM is great. It is stable and works without quirks. It does all of the things that I would typically desire from a RAID 0,1,5 setup. Administration tools are awesome and give output just as I hoped. Expand partition sizes LIVE (ext2resize needs to unmount though, that is not LVM's problem), move a file system to another physical drive, mirror partitions, spread partitions over various devices. LVM is NUTSO!
It is built into the Linux kernel past 2.4.7 (or somewhere around there), though I have heard that it was inspired from LVM for HPUX. I can't say much about this.
Understanding the concept of how LVM works can be a little hard at first, but once you get past that and then actually use it on a system, you will be totally blown away by what it does and the performance.
Here is the website for LVM
http://www.sistina.com/products_lvm.htm
I personally use Sistina LVM on a Debian Gnu/Linux system that has two IDE 60GB hard disks. I can change the sizes of partitions, move data around, move to a new hard drive on the fly, and tons of things that I don't even think I could do with the highest end of RAID controllers. As for performance, it is software RAID, but it does not have any of the typical software RAID slowness or cruft factor. I initially chose LVM as a cheap alternative to buying an IDE RAID card. Now, I don't even want an IDE RAID controller.
Men in black: "FBI! THIS IS A RAID!"
Hard drives using RAID 0: "RAID!?!?!?"
-puff of smoke-
the moral: use redundancy
...so be alert.
Each IDE controller can support up to two drives, a master and a slave. What happens if you hang two drives off one controller, and the "master" drive dies?
If it dies badly enough, the "slave" drive can go offline. Now you've got TWO drives in your array that aren't talking. There goes your redundancy.
If your purpose in using RAID is to have a system that can continue operating after a single drive failure, then you better think again before you hang two drives off any one controller.
As it points out in the Linux software RAID docs, you should only have one drive per IDE controller if you're really concerned about uptime. That would imply that "4 channel" RAID cards should only be used with a maximum of two drives, both set to "master", and no "slaves".
Note that this does not apply to SATA drives, as there isn't really a master-slave relationship with SATA -- all drives have separate cables and controller circuits. SATA drives are enumerated the same way as older drives for backwards compatibility with drivers and other software, but they are otherwise independent. (At least that's what I hear, I haven't actually seen one of these beasts yet...)
And of course none of this touches on controller failures, which is another issue. But if you are worried about losing drives and still staying up, then better take this into consideration when you design your dream storage system.
(I don't know about you guys, but I have lost several drives over the years, and not one controller...)
IDE Raid definatly has it's place. From my experience with it, IDE raid works great for lower IO services (web serving, data drives on mssql servers, dns servers, etc) however high intensity things like a mail server or anything of that nature will chew through ide drives. At the company I work for, we built an IDE webserver that worked like a charm 0 disk failures in several years.
We built a mail server (being cheap). Got 4, IBM hard drives 7200rpm, performance was great, a drive died about every 2-3 months, hard drive IO percent was about 60-70, 24/7, Tried using Maxtor, Seagate and even WD(not my choice) and all had similar results. Needless to say we changed that disk configuration quickly.
I would not suggest it to the average Joe, unless they run Windows...
/. will pay for your controller.
/dev/md0) than it did in Windows98SE/ME/2000/XP. And if anyone HAS used IDE RAID on Windows every last one of the controllers have the same problems. They CAN'T maximize the throughput without the lost of stability & reliability.
I've used IDE RAID since I purchased an Abit KT7-RAID. I've used the HighPoint HPT370 onboard as well as an Adaptec 1200A, the HPT-372/372A and now I'm using a Promise IDE RAID on an MSI KT3 Ultra.
With all the buzz words in this article for the techno geek, shuffle your buttocks to closest place to get one of these bad boys, spend that $100 bucks. If that controller benchmarks better than the SAME IDE DRIVES setup on your normal IDE controller using Linux RIAD..... well
I've went so far as to purchase 4 identical drives (as close as I could for your obssesive perfectionist boneheads). 4 Maxtor 40G ATA-100 7200 RPM drives, put them on the IDE RAID (RAID-0) controller and benchmarked them using SiSoft. Put the same drive on Mandrake 8.2/Redhat 7.3 and it showed higher drive throughput under Linux (using hdparm -Tt
For those who've lose multiple drives in IDE RAID, your victim to another reason IDE will never overtake SCSI in ANY insured business. HEAT. If you can't take the HEAT get the hell outta the real server market.
Those are the facts from someone who USES what the article "tests"
I've nothing of importance to say, now go away before I taunt you with a second sig!
What kind of hardware raid is that?
I'm using IDE raid on my home desktop right now, but I'm using software raid as opposed to a hardware controller. I have two Seagate Barracuda ATA IV 40GB hard drives hooked up as masters to my primary and secondary motherboard IDE ports. I also have a DVD-ROM hooked up as secondary slave, and a Promise Ultra133TX2 controller with a CD-RW hooked up to its first port. Both hard drives are sectioned into a 3GB primary 1st partition and a 34GB (yes, the drives are only 40GB when you're in marketing land) 2nd primary partition. Windows 2000 is installed on the first drive's 3GB partition, and redhat linux 7.3 is installed on the second drive in the same place. Both OSs share the combined 68GB RAID 0 set, which is formatted with NTFS, made from the combined second partitions. The only problem is that linux can't write to the array because NTFS write support under linux is currently "DANGEROUS" according to the driver's author and I keep important data on there. (Yes, I know about the dangers of using RAID0 and I back up regularly.) It'd sure work a whole lot better if that driver were finished, though. (hint hint, Legato Systems, Inc.) ;)
:D After a quick format with NTFS (the partiton was too big to format with FAT32), I was in business.
/dev/md0
/dev/hda2 /dev/hdc2
/etc/raidtab, ran raid0run /dev/md0, and added a line to /etc/fstab. (I read online that WinNT 4.0's software raid driver uses 64K chunks.)
;) As a bonus, I also get to keep my standard partition table as well as compatibility with non-M$ disk editing/management/recovery tools.
Getting the two OSs' software raid drivers to play nicely together was an "adventure", mostly due to Win2K's insistance on turning the disks into "dynamic disks" before letting me use its built-in RAID functionality, meaning it wanted to wipe out my old partition table, replace it with a single partition taking up the entire disk, and create a new system of partition organization inside the dummy standard partition. After a lot of reading, I found out that Windows NT 4.0 supported "stripe sets" using standard partitions, and that Windows 2000, when installed over an old copy of NT4, would support the "legacy" software RAID drive. Windows 2000 would not, however, allow me to create new legacy stripe sets for compatibility with other OSs. Stupid Micro$oft. So all I had to do was fake Win2K into thinking it had been installed over an old copy of NT4 which had been using its stripe set functionality.
The first thing I had to do was create partitions. I opened up linux fdisk and allocated 3GB on each disk to my OSs, one for linux and one for windows, and created two partitions, each one taking up the rest of the space on its disk, and set their types to 87h (NT stripe set [thanks to whoever put the L command in linux fdisk!]). After installing Windows 2000 on the first disk's first partition, I needed to get my hands on a couple of tools that didn't come with windows 2000: Windows NT 4 Disk Administrator and MS's fault tolerant disk set disaster recovery tool, FTEDIT. After spending about 6 hours searching online, I finally found a download site for FTEDIT - MS's web site says you can get it free from them, but it provides no download link. NTDA was a bit easier. Since MS service packs replace OS files, and somewhere in NT4's history a bug or problem had been found in NTDA, that file was in the service pack 6a for NT4. Service packs check to see if you're using the correct OS _after_ they decompress themselves, and they're nice enough to display an error message telling you this ("Whoops. You just wasted a whole bunch of time downloading a huge file you didn't need. Sorry!") before they delete the decompression directory. Figuring that out took a while, but snagging the executable during decompression was easy.
I ran NTDA, which populated the "missing" DISKS key in the windows registry (Win2K stores disk information in a different place from NT4), and told FTEDIT that, yes, I really did already have a software RAID 0 set on those drives, and that windows NT had died on me and I had to restore it. After a reboot, "Drive D" appeared in my computer. 68GB and unformatted. YAY!
Getting linux to see the array was much easier. I added
raiddev
raid-level 0
nr-raid-disks 2
persistent-superblock 0
chunk-size 64
device
raid-disk 0
device
raid-disk 1
to
Btw, yes, I know linux has support for MS's dynamic disk scheme. I enjoy tweaking and doing new things, even if it means days spend reading about Windows.
"So," you're probably wondering, "why did Erpo spend all that time setting up a RAID0 set (presumably for extra performance) and then go and do a stupid thing like put a DVD-ROM drive on the same ata cable as one of the disks when he has an extra ata port on his add-in controller that he's not using?" Thanks for asking. It's because Promise's bios on the Ultra133TX2 card was broken. The company "Promised" me it would allow me to boot from CD, but in reality it only will let me do so when I want to boot from a windows installation CD. Not just any windows installation CD, either. It had to be Windows 2000 Professional or XP, which I refuse to use.
It wouldn't recognize my Windows 98 SE cd, or any of my linux distros. I didn't have a choice about the DVD drive if I wanted to install linux. Just now, months after I got the card and sent promise and email, they released a bios update that claims to fix the issue. If it works I'll be moving my optical drives around. Even with the DVD drive, the performance isn't too bad - about 80MB/sec at the beginning of the disk, and it slowly drops to 50MB/sec at the end.
I'm very interested in any suggestions anyone has for LOW cost IDE RAID systems. I was considering a 3WARE 7500-8 card as they are pretty cheap these days (around $375 on pricewatch). But, it doesnt support Online Capacity Expansion (OCE).
Requirements:
1. Online capacity expansion (expansion without reformatting)
2. RAID 5 support
3. Inexpensive
4. IDE Support
I've found software that seems to support this, but if anyone has alternative advice or experience with this I'd appreciate the feedback.
RAID Reconfiguration tool
http://unthought.net/raidreconf/
What I want to see is what is the penalty for these kinds of configurations:
r aid.htm
... and based around the i80303 chip. It seems more of the higher end stuff is based around the i80303.
FibreChannel or Ultra 160 to IDE vs SCSI disks
http://www.promise.com
http://www.axus.com.tw/
The Promise enclose is based around a fixed 32MB cache and the i960 chip. Where as the axus offerings are somewhat cache configurable (64/128/256) depending
axus says the BR-8000 can do 95 MB sustained via the dual U160 interface.
So what's the real performance penalty for SCSI-SCSI for FiberChannel SCSI in this case?
Can a SCSI-SCSI do significanly more than 95MB? or is it a premium for a marginal improvement (as I suspect)?
The main reason I still go with SCSI is the reliability. I am a network admin and we *rarely* have to send back SCSI drives. IDE drives have a much higher failure rate.
:)
At home I have my OS and apps installed on a single Cheetah 15,000RPM drive (the new 15k.3). It blows away IDE RAID and completely obliterates a single IDE drive. Unfortunately buying it obliterated my wallet.
Here at work our main R&D server's been using a SCSI Mylex960 with RAID1 36GB drives. This has worked dandily for the past several years. This machine gets hit pretty had with tons of small IO, so I wouldn't consider IDE for it.
However, more recently we needed more builds/CDimage space so we picked up a Promise FastTrak100 (TX2) raid controller ($150CA) plus a couple 7200rpm 80GB Maxtors (~$150CA each), and have been living happily ever since. Now for sure we'd never put this in the main server, but for a cheap, reliable solution that gives you tons of space on a server that has only medium load, it can't be beat.
The point is, examine your needs and see what fits!
-Malloc
___________________ I want to be free()!
From the slashdot post:
I personally would love to hear any ide-raid stories that slashdotters might have.
Yeah I can't believe people didn't meditate on that sentence for a good 10 minutes before posting their own story.
The worst part about IDE RAID was I told by brother-in-law about it. RAID 0 was fine on my KT7, and I made the mistake of showing him a fast system at the time. Sure enough, he goes out and picks up a box at a local shop, puts his business on it, and wipes out file/drive/array/etc. Walking him though an important BIOS update was NOT the thing to show him.
This thanksgiving he told me how the RAID 0/1 just did not work - amazingly, he lost a drive again, and the mirrored stuff. Lets just say I never mentioned to the three other family members who have on-board RAID capabilities that it is anything other than an ATA/100 controller...
+++ UGUCAUCGUAUUUCU
Do you mean it woudn't boot the linux cds at all, or it would boot the kernel and then not mount the root, because the kernel didn't recognize the device.
If the latter, that's hardly Promises fault.. if it will recognize a bootable cd and start booting, then their claim that you can boot form it is entirely true.
Our test of the promise raid under redhat linux with the "open source" drivers (2.4.19 vanilla) compared with the 3ware product were VERY different.
I don't have the exact numbers off hand, but the 3 ware product was roughly 3 times faster at reading (raid 0+1 and raid 1). The 3ware was also faster at writing albeit the numbers were much closer. The number that DOES stick in my head was the postmark benchmark from netapp we ran. The promise did 2500 files, from 2 to 200k with 500 operations in about 35 seconds. The 3ware product did the same in 12.
The moral of the story is TEST TEST TEST, these types of articles only give you an idea. Promise worked great for me personally in several applications. After testing it for a production machine at work, we went with the 3ware because the promise did not perform well for our application. Test for youself, or forever be dissapointed.
Cluge
"Science is about ego as much as it is about discovery and truth " - I said it, so sue me.
This "Crap" is because, in ages past, a bunch of smaller drives were WAY cheaper than one large drive.
IT was a way to get large storage space out of small drives, originally. The redundancy issues are there to reduce the failure rate of the array to something matching a single drive (as opposed to say, 20 times as likely if you add 20 drives)
I was doing some cross-platform admin work in a mid-end pre-press house and they were still back in the 80's as far as their networking and storage went. Ended up putting together a proposal to build a server for them, and the best solution included the Promise FastTrak ATA-100 Controller.
The server was built around an Abit BP-6 with dual C-400s, 512Mb RAM and 160G RAID 0+1 storage running WinNT4.0 Server (had to talk Appletalk). I installed it, a new 24 port 10/100 switch and re-cabled the entire network in December 2000. As far as I know, it still runs just fine and is serving files great. It also acts as an Anonymous FTP server, DHCP and print spooler for 5 printers.
Kevin
What about RAM-RAID?
That's what I do when I want a faster computer.
I recently built a box for myself, pretty good should last 2-3 years (or until DOOM III comes out...o wait...nevermind)
anyway I'm using an Asus P4B533-E mobo with built in promise fasttrack 133 IDE Raid. I have 2 Western Digital Caviar 80Gb HDs striped, and so far, the write/read performance is loads faster than any computer i have ever used. No problems (although it does take about twice as long to reboot) Anything REALLY important i back up on cds, sometimes multiple floppies and on my webspace, eventually i will probably buy 2 more drives and make it a RAID-10 (read: not 0+1, but similar) because backing up some of my media takes 20 cds a month (that piles up quick).
Maybe its time for me to get a DVD-Burner?
In the article, it states:
RAID 10 can survive a second failure if it occurs in a different mirror group than the first failed drive. However, if two drives fail within the same mirror group, all you're left with is half of a striped array, which is useless.
That doesn't sound right to me. If you have a mirrored pair of striped sets with parity, you should be able to survive a either a failure in both sets (essentially losing parity in both, but not data), OR two drives failing in one set. If you lose two drives in one set, it's basically the same as losing one half of a mirrored pair of drives; you're left with the other half of the mirror - that's the whole point of mirroring.
In fact, you should be able to get through a triple failure - two drives in one set (which kills that set) plus one drive in the other set (so you lose parity).
Confusing, I know, but it seems like a significant oversight on the part of the author, when he's explaining the technology.
-HL
Of course it's not gonna come even close. Every person (or at least people vaguely familiar with good ol' scuzzy) knows of SCSI's advantages. Current IDE and SCSI aren't even close in performance, sure... but they're also not even close when it comes to price. $70 for a 40GB IDE drive... $140 for a 18.4GB SCSI drive. Gee, let me think. Unless your PHB is picking up the tab...
RAID 5 in software can be dangerous. If a parity write fails (disk/system dies), you'll likely have data corruption and not even know it. Best to trust reliable hardware to do the XORs.
Then again, a RAID _card_ may not help here, since the disks are at the mercy of the system power. Best to use a real array, if you have the bucks.
WTF are you talking about, they are almost ALL under 2%! Only the promise controller goes for 7%.
Do you mean it woudn't boot the linux cds at all, or it would boot the kernel and then not mount the root, because the kernel didn't recognize the device[?]
It wouldn't boot the linux cds at all. The card's bios would not detect that the CD was bootable, that is it would totally miss the ElTorito-compatible 2.88MB floppy disk image on the cd if the CD was not a Windows 2000 Professional or XP install disc. Once I moved the DVD drive over onto a motherboard controller, redhat 7.3 installed perfectly. It even loaded linux's pdc202xx driver, detected my CD-RW (which was still on the controller with the buggy bios), and configured ide-scsi emulation so that cd burning would work.
Wow. I'm surprised at all these stories of hard drives going south. I've never had a hard drive die, in many years of use. I still have a badly abused 10 year old 210MB harddrive that still works great. Ditto for the slightly younger, but still badly abused 1GB drive. That said, I don't worry about data loss anymore. Data is far too important to keep on your working machines. Thus, I've got a Linux server where I keep all my data. I've gotten into the habit of putting everything under subversion (it's alpha software, but remarkably stable) so I get everything version controlled for free. Throw in the occasional backup, and not only is my data (all versions of it) acessible anywhere I've got an internet connection, but it should stay accessible in the event that I kill my main machine by abusing it like I do.
A deep unwavering belief is a sure sign you're missing something...
To all those using RAID personal computer: Is it really worth the noise? That grating whine of hd spindle is the most annoying sound pcs make.
Obviously, software RAID is better than nothing, but there's a reason hardware RAID is preferred.
What does RAID do?
Protect against hardware failure.
Where is the RAID configuration stored?
On software, it's stored on the very same hardware that you're anticipating failure on.
Everyone say it with me: there are lies, damn lies, and benchmarks.
Free messageboards and more! Your girlfriend's seen myWang
Why the hell does a *workstation* need RAID or SCSI? A single IDE drive should be fine! Unless they've went and changed the nature of what a workstation is again.
As for small scale servers, guess what? They're already running IDE, RAID or no. There's a lot of servers out there that just don't need the speed and general beauty of SCSI - it isn't worth the extra expense.
Put it this way. If I was designing a multi-forum pirate news site called ArrrDot, I'd go with SCSI due to the insane disk I/O I'd need.
If I was designing a multi-forum underwater basket weaving news site called SlashBoredom, I'd probably stick with IDE. I might regret it when someone posts the site to Slashdot, but for day to day operations, spending more money on SCSI doesn't make sense. There's not that many underwater basket weavers out there.
First off, they've failed to note that some of their contestants are in fact just IDE controllers, with the RAID functionality implemented in the software driver (WinRAID, like WinModems), whereas others are Hardware. I don't know all four products well, so I'm unsure on at least one of them as to which are which.
They tested CPU utilization, and seperately various speed tests, but never a comprehensive "loaded system" test. As expected they ranked the Adaptec (a true hardware RAID) lowest, while ranking the WinRAID's higher. This couldn't be further from the real truth. Sure, the idle P4 cpu does a great job of fast software RAID compared to the embedded RAID ASIC on Adaptec's card. However, if you had a heavily loaded server machine, where the processors were loaded down doing other things (say SSL-encrypting for an secure web server), the machine with the Adaptec would trounce the others, as the RAID processing speed will not decrease while your applications are using most of the CPU (or depending on the device driver's pre-emptability, it could be the other way around, that the CPU simply wouldn't be as available to your CPU-hungy SSL server as it's busy with the RAID).
11*43+456^2
In my opinion, IDE RAID is near worthless. In most cases, if your data is valuable, or you need performance, you will lay out the few extra hundred dollars for SCSI. The performance is far better (10-15,000 RPM these days), it is far more scalable, with 32 devices a channel, as opposed to 2, and the protocall is better. IDE is a clumbsy format that shouldn't have been scaled up to the sizes and speeds that it has.
The Seagate barracuda IV had a problem when connected in RAID-0; it actually performed worse than as a single drive. There was a problem with caching on them that made RAID controllers gag. There is a firmware fix; you can contact Seagate and they will replace your drives for you! I currently have 5 Barracuda IV drives connected to my highpoint RAID controller (abit at7) but they are running as single; I use the speed advantage of each being on their own IDE channel.
Now Soundgarden was headed for the bottom in track #4, Mailman. Had the clip lasted just a little longer, you'd know that they were riding you all the way, just like the Party back in Soviet Russia.
Hopefully, Chris Cornell hasn't died like Layne Staley and my Alice in Chains post did earlier today...
Get thee glass eyes, and, like a scurvy politician, seem to see things thou dost not.--King Lear
I've used a couple ide based raid cards in the last few years, starting off with a motherboard that has a promise chip, then a promise card and finally a 3ware.
The onboard solution was full of problems simply due to the fact that bios updates for it were dependent on abit and they didnt keep up to date, the technology seemed fine so i went ahead and grabbed a 4 drive promise pci card.
The 4 drive promise card worked well, a few driver issues but all in all a good card for the price, i ran 4 maxtor 40 gigs in raid 0. One died of course and bingo, i'm fucked. Rma'ed it and ended up distributing the 4 drives around my house for various use.
My current raid is 6 of the seagate barracuda IV 80 gig's in raid 5 with a 3ware escalade 7810, I've been extremely happy with this card. Having dealt with mylex and adaptec scsi raid cards in workservers I was extremely pleased to use the card, the software is all web based which is nice for remote administration and is very minimal, no bloatware that you often see seeping into drivers these days. ( Here's copy of my raid details page from the other day when i was rebuilding after my system crashed due to a loose drive power cable ).
About 6 months ago my raid crashed and no matter what I couldnt seem to initialize it without having to delete it all and restart, I figured i might as well contact 3ware first, went to their site and used their support page, the next day first thing in the morning i got a call from Angie at 3ware, I talked to her and on her advice upgraded the bios and software versions to the latest which had apparently implemented some bug fixes, she also gave me her number so that in the event it didnt help she could send me some debug utilities to get more information about what happened and how they could solve it. The bios upgrade worked thankfully and I never had to return the call. Honestly, I didnt expect that level of customer support and hands down I would recommend a 3ware card to anyone, the performance is lower for writes but their support more then makes up for it imho.
FYI, I've also used my 3ware card in linux and freebsd and the management software worked just as well as in windows, props to the 3ware team for that.
Quirky? How about downright wrong?
http://www.fornax.net/platforms/slave-i_raid/
The only IDE RAID controller that blows SCSI RAID away in size and data burst rate are 3ware's controllers. There are architectural limitations to IDE, and 3ware knows every shortcoming and compensates for it. One drive per channel and controller chips on every crevice.
I have 8- 120 Gig drives in a 0+1 array. You would never have seen such a thing on SCSI. And best of all- guess what. It costs jack shit.. It's beer money. You need to take out a mortgage to get the same results from a SCSI array. Oh, and it runs on linux since they open sourced their drivers quite awhile back. IDE RAID is a good tool to consider using if you need massive storage and aren't a compulsive spender.
SCSI is great architecturally, but who are you people kidding with 800 dollar shit for space drives. How much better is your data gonna be sitting there wasting all that money. Honestly, if my hammer cost 800 bucks I'd put it through my skull. It's just a tool. Use what works.
I have a Promise FastTrack 100 TX 4 Raid controller purchased twelve months ago.
To cut a very long story short their driver support absolutely sucks. To this day there is still not a Windows XP driver for the card, and the windows 2000 driver I was forced to use took a lot of hassle to get working. They realised a new version of the Win 2000 driver that just hard locks my machine. Apparently this type of problem is fairly common with Promise controllers.
The worst part is that numerous emails to their support dept went unanswered, finally got one back (only 3 months after sending it!) [yes this is email, not snail mail] to say buy the new card.
I made a promise not to buy another one of there crappy cards again.
Games Programmer And Designer
I work for an ISP in the Midwest. We were severely limited on budget when we built our new data center. The CIO at the time insisted that we use the Promise Fastrak IDE raid cards (Under Redhat Linux doing Raid 1) Within a year; we had two separate system failures linked to the Raid controller. Both times, a single drive would fail and then Fastrak card would kill the second drive while attempting to recover and then die itself. Frustrating to no end, especially in light of the system with an Adaptec raid card doing Raid 5 that had been running non stop for approximately 3 years. Wiser heads now prevail and we are moving everything over to Adaptec SCSI raid cards (2100S to be exact)
What DDS drive / software do you use that is audio enabled? The only time I've heard of something like this is on an SGI workstation.
I got a Promise FastTrak 4X or something... and two 7200RPM 30GB drives to go with it. Got home, hooked everything together. I wanted RAID-0. I put in my Windows XP CD (I hadn't discovered the World of Linux back then) and the freakin' thing would go through all the install but fail after the first reboot. I spent about 4 nights solid trying to get the damned drivers to work. Emailed their tech support; they told me to use different drivers. Which didn't work.
After 4 days, I installed Windows 2000 and did an upgrade to XP. Surprisingly, it worked. And what was the speed difference? Miniscule. A bit noticeable when playing games, but that's it. And the damned thing added 45 seconds to my boot time! I mean, really!
Linux drivers were binary-only (I think that's changed now) and Linux recognized it as a normal IDE controller.
I returned the card and now have the two 30-gig drives on my motherboard's integrated IDE controller. I don't plan on getting RAID any time in the future (except perhaps when I'm very rich and go for a SCSI RAID-5 or something). None of my programs need it; and that extra 45 seconds every boot is painful, even if it's only once every two months. Considering my entire system's boot time is 30 seconds and the biggest program loads in 2 (well, barring games) I don't see myself needing RAID for personal use.
Debian Raid1
Using hdparm(1) to tune up the disks, gets a sustained 35 MByte/s reconstruction. Very good as a small-medium environment file/print/mail/proxy/dns server.
I setup Software IDE RAID in linux about 6 months ago. I had a fast 40 gig drive (IBM 60GXP) but I heard about a lot of failures with them, so I took my old WD 20 gig, slow piece of crap and mirrored my 20 gig linux partition on it (I wasn't worried about losing my Windows Partition). The performance hit was very mild, surprisingly. The Software RAID was smart enough to put most of the load on the faster drive, at least for reading. Writing was the only thing that took a bit longer, but only about 10%. A few reading functions were actually a tiny bit faster, according to the Bonnie benchmark. Overall, the piece of mind was worth it. . . until the county Sheriff stole my computers....
We ended up going with the Promise UltraTrak SX8000, which is an external RAID cabinet that holds up to 8 IDE drives and connects up to the host computer via SCSI. We then got 8 120GB Western Digital drives for around 150$ each. The RAID set up quickly, and within an hour we had a formatted 7-drive RAID 5 array with a hotspare for if things went badly.
The cabinet has, in the 4 months since installation, given us zero problems, and worked flawlessly, with quick transfer rates, and extremely easy setup. Considering the price compared to an equivalent SCSI system, we feel that we got 90% of the value of a SCSI system (the only difference being that IDE drives break sooner than SCSI drives, and that SCSI drives are moderately faster, both of which weren't quite necessary for us.)
If your system contains mission-critical data, go the more expensive route and get a full SCSI raid system with multiple hotspares and pay a guy to sit in a corner and maintain it. If, like us, you just need a large amount of very-reliable storage without much hassle, go the IDE RAID route. It's working great for us.
I've got 5 servers (one is an Exchange 2000 server) at a school with about 200 users. All servers are running some form of promise ATA raid 1 setup for boot drives and some also use an ATA raid 1 for their data drives. The file server and mail server use Adaptec 2400A raid controllers with four 100GB drives in a RAID 5 configuration. All hard drives are western digital 7200 RPM drives.
No one complains about speed issues. Everything seems to work very well.....at a fraction of the cost of SCSI.
I love the look on visitors' faces when they see our servers have 300-600GB of available storage...for very little cost. (Backing up all that data still requires SCSI tape arrays...not cheap.)
I've had a couple of drives tank on me here and there, but no data loss yet...just replaced the failed drive...rebooted and in about 20 minutes the array was completely rebuilt.
I am a fan of SCSI (got plenty of SCSI raid at my house) but when you've got to stay under a budget, you can't beat ATA raid.
-ted
From the Slashdot story: "I personally would love to hear any ide-raid stories that slashdotters might have." I also would like to hear about this.
Here's my story: I have extensive experience with Promise controllers. An IDE mirror makes data reads faster. If you are about to do a possibly damaging operation, it is good to break the mirror, pull out one of the hard drives, and do the operation on the other drive only. Then, when craziness happens, the other drive is a complete backup.
A mirroring controller is a convenient way to make a Windows XP operating system hard drive clone. Windows XP prevents this; normally third-party software that runs under DOS is needed to make a useable full hard drive backup. See the section "Backup Problems: Windows XP cannot copy some of its own files" in the article Windows XP Shows the Direction Microsoft is Going. (The article was updated today. To all those who have read the article, sorry for the previously poor wording of the section "Hidden Connections". Expect further improvements later in this section later.)
But Promise controllers are quirky. Sometimes things go wrong, and there is no explanation available from Promise. Promise tech support is surprisingly ignorant of the issues. The setup is quirky; it is difficult to train a non-technical person to deal with the controller's interface.
Mirrors are a GREAT idea, but Promise is un-promising. That's my opinion. I'm looking for another supplier, so I want to hear other's stories.
I've set up an IDE Raid array in my machines three times. The first time with an Adaptec 1200a and two Maxtor DX740's. I assumed that if I striped them I would get a significant performance increase. After running benchmarks with HD Tach and SiSoft before and after I was getting worse performance and high cpu utilization on the striped set than on a single drive. I went through every option with Adaptec Tech Support (which were great) to no avail. I tried the same drives on a Promise Fastrak and got slightly better results but I kept getting write-behind errors and corruption. The only success I've had was using it in a Raid 1 configuration for redundancy. I am definately going to go for a SCSI solution when I get up the extra cash.
True, but you'd also think that after 130 graphs and countless pages of whining about how the RocketRaid 133 had only two controllers he would've realized that the Highpoint 404 would've been a more appropriate choice for this benchmark!
(Still and all, the "404" is a lousy name for anything that's intended to make sure your data can always be found)
Am I the only one who heard Roxette to sing "I'm gonna get blitzed for some sex"?
I've passed this feedback onto the author of the ide raid-roundup - i figured i might as well post it here too.
I just thought i'd share some of my experiences with promise support.
Frankly, they have been terrible. I would not voluntarily buy another promise product again at this stage based on my experience with them.
I have been attempting to get support for the Promise FastTrack which is a popular embedded raid controller option, under Linux.
Promise indeed "support" RedHat but do so with a binary only, closed source module that in the end turns out to be useless.
Promise hard code a supported kernel version for this driver such that you can run it under say RedHat 7.3, but only the initial 2.4.18-3 kernel, which has a number of critical bugs which have been addressed in later (errata) kernel updates.
Needless to say, promise's driver will not run on any later kernel or at least they are unwilling to answer questions on how to do this.
A comparable analogy would be if they had released Windows XP drivers and then your hard drive failed to work if you installed a hot fix or a service pack because the driver is keyed to only the specific intial installed released of XP. Promise don't treat windows users this way, so why do they do this for linux users ?
I've managed to get two responses out of their support, none of which will address my problem - support the hardware under linux by releasing the source or provide updated kernel drivers for the released kernel images that will actually work.
In terms of driver support for Linux/FreeBSD, 3ware wins hands down in this group.
regards,
-jason
No kidding. Especially with a cost-capacity ratio. On SCSI, to get anywhere near the large amount of IDE capacity, you need multiple drives. And of course, drives in SCSI are not cheap by any stretch of the imagination.
I'd rather shell out for a few 80/120GB drives than be nickel&dimes by a lot of smaller ones, although in the case of SCSI it's more like quarter-and-loonied (dollared for you Americans)
I have a 3ware Escalade 7500-8 in my home server with 6x120G drives in RAID-5.
Cabling was sort of messy (but only because it's a plain big tower case and not a double-wide fileserver thing) but other than that I plugged it in and it worked. I have two 3ware hot-swap RAID enclosures that allowed me to install the six drives in four 5"1/4 bays - they're great, and extremely well-built, albeit a tad on the expensive side for just some cleverly twisted metal. (ie no electronics) And hot swap is great. Pull a drive out while the computer's on and the alarm goes off, put it back, and it'll shut up. An awesome toy.
Performance is more than enough for me, I just use it for archival puproses.
Two things I dislike about the card:
1. Array builds can only happen in the BIOS. If you create a new, really big array you can be prepared to spend a couple of hours staring at the screen.
2. Linux software support is too good. Matter of fact, better than Windows support. This is a bummer because I run W2K on the box.
I have a 486 running both onboard IDE and a Promise Ultra66 controller. I have not had the courage to make them RAID anything yet. The reason is that I have a 4.3GB HD with bad sectors. Easy enough to manage. I partitioned 1 GB off of the drive because it was the first gigabyte and had bad sectors scattered throughout. I am running Windows 2000 Advanced Server on this machine. The specs are: 2.5 GB HD, 4.3GB HD, 8x CD-ROM, 64MB RAM, Matrox Millenium video card, 3Com Fast EtherLink XL, PnP BIOS (quite nice for a 486), AMD 5x86 (souped up 486 chip.....) at 160MHz, PCI V.90 modem, SB AWE64 sound card. To top it all off I'm actually running Windows 2000 Advanced server on it. I used to run Caldera OpenLinux 2.2 on it, then went to NT4 (a buddy of mine who LEGALLY bought it with Proxy Server 2) then upgraded to Win2k. Has anyone else tried RAID with drives with bad sectors??? anyone soup up their 486??
I recently compiled the 2.4.20 kernel to get my Promise PDC20276 controller and RAIDed partitions working properly under linux.
Linux could see the two seperate drives i have (one on either of the promise ide connectors) but not the one raid drive.
According to the help in the kernel, the linux drivers ignore the promise's bios interface (which presents to two drives as one raided) saying that it's too slow. It sees the two drives seperately and then combines them using software - the module in the kernel (the ideraid.o module).
So it all comes down to this-
If you want to use the raid drives as one logical drive in both windows and linux you should make the raid set using the promise bios, then use the latest linux kernel with the beta promise support.
Alternatively if you only use Linux - then don't worry about doing the RAID in the promise bios. Set them up as two seperate drives and use Linuxes' pure software RAID - it's more flexible (can use miss-matched drives, can mix and match mirror and stripe) and performance will be the same.
If you use software raid in windows you can't have your root partition (usually C:\ drive) on it (but can mirror it later on) BUT you need to make the disks windows dynamic disks (their version of BSD disk slices) which might kill linux use of the drive (although linux has beta/expermental support for windows dynamic disks)
I just saw this:
I'm getting one for my house with 200GB drives....they also make an 8 drive version....mmm....
-ted
software raid, 2 disks on a promise ultra100 tx2 /usr, /tmp) that can be easilly replaced
one for each channel. ONe is a 80 gig, the other
is a 40 gig. Some of the non-important files
(such as
are on partitions that aren't mirrored. The important data are on partitions that are mirrored on both. Software raid, level 1 (mirroring). Its what it is: Cheap insurance. Works great for me.
Windows 2000 Advanced server on a 486? Are you one of those guys that enjoys pain?
Seriously...if you want the PR 233 and the AT motherboard (with AGP...maybe even 64 or 128 MB memory), reply to this and we'll figure out the shipping address thing. I can't let you suffer like this man!
-ted
dual xeon 2g
2g ram
16 120g 5400rpm maxtor drives across 2 7500-8 3ware cards
2 gigabit network cards
We were pretty much forced into a xeon solution because it was the only motherboard I could find that had the number of PCI slots and busses that I felt comfortable with. The two 3ware cards sit on one 64/66 bus and the main NIC sits on another bus. I use one array for the live data set and one for the backups. (LVM + ReiserFS in linux is the best thing since sliced bread)
Its really neat being able to increase the size of your filesystems on the fly, or just create another 80g lv to hold some temp data. For backups, I just snapshot the lv I want to back up, create a new lv on the second card and just dd the data over.
I have a boatload of bonnie++ numbers that I got when I was testing it out. If anyone is interested I can post them. We settled on RAID-5 on each controller because even though we took a performance hit, the network was still the bottleneck.
All this in 4u of space. Oh, if you do build a mega-server like this, TAKE TIME TO CALCULATE POWER REQUIREMENTS! The power supply that was going to be supplied with the case was WAY undersized for spinning up all the disks at once. I had to special order a supply (zippy.com R3G-6650P, their page is slow and hard to navigate, be forewarned)
I'm tired, I think I'll go to bed. If you want more info just reply to this and I'll get to it in the morning.
-paul
HERE
I can vouch for the stability and speed of one pre-built IDE-RAID Product. RaidZone OpenNAS They use a special raid controller that allows increased IDE Raid speed, and IDE Hot-Swap capability. This one had a total of 1.2TB (plus hot spare). The project involved a unique (to this project) Application that required a proprietary Database System that could only run on an MS Win2k Server, therefore we didnt fully utilize the sytems capabilities. It served as a file server for images (ranging in size from 60kB to well over 100MB per image) (and as I recall the Image Store is now around 500GB), MS SQL Server, confidential proprietary DB system (indexed images, among other things), and several small services (and the chunky MS GUI). It even has a 900GB native backup system attached. The load ranges. Since we put it into action early 2002 it hasn't missed a beat. I would recommend it highly for most applications, though there does come a time when higher speed drives are needed.
Why hasn't the ArcoIDE solution caught on like wildfire? It provides mirrored disk capability with absolutely no visibility to even the motherboard, much less the OS. I've been running it for years and it's great. Mine is the PCI slot model that simply uses the slot to get power to the card. One IDE cable from the motherboard to the card, two cables to the two hard drives.
And there's all sorts of alarming options -- LED's on the card, LED's on a front panel bezel, audible screech, Form C contacts for you industry types ...
I don't get it.
One simple rule for its versus it's
I ran a RH 7.1 box with 4 100GB IDE drives in a software RAID5 configuration (300GB data, 100GB stripe) for a few months. It seemed pretty stable until one day I walked in and saw a kernel panic dump on the monitor. I crossed my fingers and rebooted to find that the drives weren't synchronized (one of them had a stale event count). I searched all the documentation I could find but I could not figure out how to re-sync the array without losing data. (The system wouldn't boot with the bad drive removed either.) In desperation, I called RedHat. They said "Software RAID support requires a higher level service contract.". They wanted $3,000.00 before they would even talk with me. I offered them $100.00/hour but they refused my business.
I gave up and re-initialized the whole array (after making some hardware changes). Fortunately I had a fairly recent backup. Total down time was about 3 days.
At my house I have two dual drive mirrors. (I started with one and added the second later, hence the lack of mirrored stripes or striped mirrors.) I chose this method because I generally don't have problems with virii/trojans and am aware of the consequences of deleting the wrong file(s). I was more interested in protection from physical drive failure.
The actual setup is an older Socket 7 machine, two Promise PCI IDE cards (one ATA-66 and one ATA-133), and Linux software mirroring.
My former workplace uses two of the Arena Indy 2400 units on their backup server. This box stores data from 150+ other machines on their network that can be restored by their system administrator, so raw access speed isn't that important (i.e. it isn't a general purpose file server). Lots of space, however, is important. They have enough space (~2.4 TB) to store a copy of each machines drive and about two weeks of changes.
That machine is a single processor P3 running Linux. The filesystems are reiserfs since their favorite distribution (Slackware) supported installing to reiser (and not yet ext3) at the time the box was built.
Wouldn't your (raid 3) array performance be bounded by the overhead of parity calculation? Wouldn't the write performance be made worse by restricting all parity writes to one hard drive? It seems by distributing parity among all drives raid 5 still carries the parity calculation overhead, but is faster at writing the parity simultaneously to all drives. Am I just splitting hairs here?
-ted
I can never figure out why massive reviews like this never compare the speed of the hardware controllers to NT/2000 or Linux Software RAID.
Informal tests I have done show the software RAID compares very favorably to IDE hardware based solutions.
So I was surprised reading the review to see the Adaptec and 3Ware neck and neck in the RAID 5 area. 3Ware's usually have no competition in RAID-5 since their firmware and HW rock.
Then I found out WHY they were so close:
The 3Ware cards are 64-bit cards while the Adaptec's are only 32-bit. 3Ware cards can hit 70MB/sec writing and over 150MBsec reading with 8HD's! If they ever get to 66MHz, I expect their performance to go even higher.
If you want to see better benchmarks that fit with reality, check out the XBit Labs Review
Top Most Bizarre/Disturbing Error Messages
Utilizing eBay and a few vendors that I dug around for, I was able to assemble a blazingly fast fibre channel RAID system for home for around $500. If you take a look at http://www.nuxx.net/gallery/fibrechannel you can see the assembly of the box. There are also benchmarks detailing the RAID 5 array bursting to >160MB/sec (image at http://www.nuxx.net/gallery/fc_benchmarks/aad).
The box is set up as follows:
o Mylex eXtremeRAID 3000 ($200 via eBay)
o Crucial 256MB DIMM for Cache (~$50 from Crucial)
o 4 x Seagate ST39102FC 9GB 10,000 RPM drives ($9/ea on eBay)
o Venus-brand 4-disk external enclosure (~$35 on eBay)
o Custom made FC-AL backplane for disks (~$200 from a site I can't remember at this time)
o 35m FC-AL cable (HSSDCDB9) (~$40 for two on eBay)
The best part? The box is located in my basement, so I have this incredibly fast disk disk access, with no noise and no extra heat inside my case. That also allows me to cool the case more efficiently. Sure, IDE RAID may be cheaper, but the performance, per-disk, coupled with the reduced noise in my office and the reduced heat in the case is a big plus. Also, I might eventually pick up a second backplane for another four disks and do RAID 0+1. Since each channel is capable of 100MB/sec (without caching), the use of a set created across two channels would be amazing.
I saw the LVM stuff, and was going to use it on a new install with two 80GB drives (and a separate boot/OS drive). But I wasn't sure about redundancy. How does it handle that? I mean, can I use a "RAID1-esque" pairing with it? Does it do that already? From what I saw, the LVM stuff is really just a move away from BIOS partitions (and terribly cool for those times when /usr/local starts to run low). I'm looking for data security through redundancy, but like the features LVM has.
The machine I'm talking about is a samba/NFS fileserver on a 100mbit (full-duplex, switched) LAN, so I'm not all that concerned about speed as much as I am about recoverability -- as long as disk reads and writes are faster than what comes in over the network I'm plenty happy. The ability to grow "partitions" would be a wonderful added bonus.
It's certainly interesting stuff...
-B
Ash and Hickory, straight-grained and true, make excellent bludgeons, dandy for the cudgeling of vegetarians.
I got this card along with 4 Maxtor 20 gig 7200 RPM drives, and 256 MB DIMM for cache. Before installing into small file server, tested it out to see what the raid 0 performance would be across these 4 drives. I was amazed that it seemed actually slower than my regular 15K SCSI drive. Not quite the performance I was expecting. It does do well with large writes, just access times seem like a regular 5400 RPM IDE drive. The card was like $175, and does RAID 0, 1, and 5. Can also do 0+1. The PAM utility (Promise Array Monitor) isn't quite intuitive, but gets the job done. It's a good card for simple reliability - just not as fast as one would think...
Thanks.
today is spelling optional day.
Sure, it's a parallel cable. But only one device can communicate with the host at once. It's a lot harder to squeeze 100% of the theoretical bandwidth with 8 slow devices than with 4 fast ones. So the parent's comment was mostly accurate.
THIS THING CAN TURN ON A DIME, MACROSSZERO STYLE ALSO FUCK BETA, ~NYORON
ide to scsi adapters.
Get those nice big fast IDE drives, slap a scsi adapter on the back, and stick it in a hot-swapping rackmount unit with a single scsi port on the back to run into whatever server needs a bit more space.
We, at my work (ISP), use them at for our news servers, and whenever we feel like it's time to add another terabyte to increase article retention (read: customer's precious warez/pr0n), we throw on another one of these bad boys:
12x 160GB Maxtor IDE fast spin drives
Hot-swappable by design
dual redundant p/s and so on and so forth.
All you need is your scsi raid controller, which I admit ain't cheap, but you can make up for it with those oh so cheap IDE drives and creative use of IDs and LUNs
They're all hooked up to a 3ware Escalade 7500-12 card, RAID5, with a hot spare. Application is storage of large amounts of raw digital images 7-8MB each.
Been going for a few weeks now, no problems, 2.4.19 kernel's built in drivers lights the array right up as sda1.
I would show you more but I'm ssh'd in and the power just went out. The 300VA ups running this box while I'm testing it probably just let its smoke out. Doh.
Anyway I like it. If its not fried.
--------------------------------
Not all who wander, are lost.
At our business we use the 3Ware 7500 with a 3x 120GB (1200JB) Western Digital RAID5 configuration on Linux RedHat 7.2. The machine is a dual P3/1GHz on a SuperMicro 370DER motherboard. We use the machine as our primary file server/compilation box, so data integrity and fast failure resolution is critical.
The cited benchmark page has excellent information (130 graphs!), and it confirms my first hand experiences of everyday use of the 3Ware 7500. The read times in RAID5 are outstanding, but there are sometimes significant delays on file creation. In addition, it seems that IO is single piped, or serialized on writes at times.
Since the 3Ware 7500 is based on an FPGA (Field Programmable Gate Array, (see http://www.xilinx.com/), with LOTS of extra ECC processing power, the problem with writes is not likely to be related to the "CPU", but rather part of the internal IO structure of the card. I hope it is amenable to correction with a microcode upgrade to the FPGA, but it may be related to the memory architecture of the card.
Overall, I'm pretty happy with the controller, especially the ease of installation and rebuild time. I have high confidence in the data integrity, and the price is good. I also recommend the drive cage that 3Ware sells, even though it is expensive at $200, it's well worth it because it fits 3 drives in the space of 2 5 1/4 inch bays, and it is hot swap proven.
I'd like to see a shoot-out between the 3Ware and some SCSI/Fibre Channel RAID 5 configurations!!!
Comments?
---- Luke "To boldly go where no one has gone before..."
I always do..
Was running RedHat, using generic I2O drivers, not the Promise drivers, no big deal, though performance wasn't probably as good as it could have been. My main concern was data integrity, not performance.
So anyway, this system is in my bedroom (which sounds like a busy airport, but i like it) and one night I woke up because it was beeping, which means something no worky. I had noticed a few days earlier that I was getting pauses during reads from the disk, which was unusual -- I was going to put the Promise drivers on and see if that made any difference. But then the beeping came. So go in, system still running fine, shutdown and go into the raid bios. It says it's trying to rebuild, and it's failing, it isn't saying it has a bad disk, and even if it did say that, it couldn't tell me which disk because they didn't think that info would be important when they made the product. So, shut it down and pull out each disk, run WD diags on them. Found that one of the disks, during the sector walks, was pausing every now and then for no good reason. Fine, put the working disks back in with a new disk so the array can rebuild. Array rebuilding.. Array fails during rebuild.. WTF? I tried about a dozen configurations and the thing simply wouldn't rebuild. Well, @#$&*!, I'll just rebuild everything from scratch, so back up the data elsewhere on the network. Boot up (i boot from the raid array, I know bad bad), kernel panic after about 3 minutes (it's still trying to rebuild in the background -- and failing, why the panic though i have no idea, might have been the i2o drivers). Take out the new (blank) drive, no more kernel panic (because no more trying to rebuild). So... copy data to network. Put new OS on (this time Gentoo) on a 4GB hard drive for booting only. Rebuild the raid array (which takes f-o-r-e-v-e-r). Copy data back, install apps, lalala, big pain in the ass.
So, bad disk and bad raid controller IMHO. Should have been able to rebuild. However, I did not lose any data. New system has been up and running now with no further problems. Also, performance seems better when using the actual Promise driver, go figure.
Stupider like a fox! - H.S.
We've been running a Tyan K7 Thunder dual processor, 2 gigs RAM (I know - overkill, but the machine was built to accomodate its life later as something else).
IDE RAID was twofold:
(1) Adaptec 1200A and 2 mirrored drives - 30 gig IBM Deskstar 7200 GXP for OS
(1) Adaptec 2400A and 5 RAID5 7200 75 gig IBM Deskstar 7200 GXP for storage
It is a very tall tower, but everything actually fits in it OK, though the cables would be better served with the smaller round kind.
We've been using it as our major file server for almost a year and a half, no data loss, no drive failures.There might be some gasps at the drives we chose, but I can tell you that keeping 7200 RPM drives cool is the secret to their longevity, and the file server is in a very cold room.
Getting Windows 4.x Server installed on it was a bit challenging, but after a quick call to tech support (and finding out it was a silly Windows issue), everything was very smooth. Eventually it will host Linux or Win2k; yet to be decided.
We've never had a drive go down, but just LAST weekend the motherboard and power supply went south. We went to Fry's electronics and picked up a newer Tyan K7 Tiger (lower end board) and inserted the same RAM and Processors - attached all the drives and fired it up and it ran perfectly!
We've been very fortunate with this system. I think that IDE RAID has been just as good for us as SCSI RAID, and it got us a very large file server at a very inexpensive cost!
HOWEVER, I caution any of you wanting to skimp on $$ that the 4 hour service warranty you might pay from Dell or HP/Compaq, or IBM will cover your ass in a tight situation. Don't base your career on store-bought parts! Also, we still don't have spare drives for our system, which is just plain dumb. Spend the extra $$ and get a drive or two.
Furthermore, consider a backup power supply, because I've found those seem to be degrading in quality over time (I've replaced two at home, and a few at work).
Finally, our original plan was to replace all the IDE drives in the RAID5 array every year. Unfortunately, this hasn't happened. I think that this would be a wise move for anyone building such a system.
Cheers,
&J
We have quite a few external 8-bay IDE-RAID units ( called RAIDKing .. ~$6K w/ 8 disks ) with a LVD interface that are just horrible when things go wrong with the disks. I beleive they use some king of HighPoint contoller inside.
.. the red LED on the tray will light up and the units beeps notifying you to put in a new disk to rebuild but we have had some cases recently where you get SCSI errors but no indication from the unit!! If you try to re-build the RAID .. it will fail with no indication which drive had a problem! Then you have to try and rebuilt starting with three disks and keep adding one as you go until the re-build fails. Very frustrating!.!.!
... you get junk shit. ... your shit don't stink.
You can choose a variety of configuration options RAID 5(with or without a hot spare), 3, 0, and 0+1 but there is not reporting or monitoring on or about the unit so trying to debug is useless
Normally when the disks have a problem
I say the cost break compared to a high end RAID system such as LSI, DataDirect Networks(rocks), or SGI is not worth the headaches with cheap low-end systems.
You buy junk shit
You buy quality shit
hehehehe, it way to late for me
I had to replace one of those IBM GXE drives that have been falling over all over the place in an adaptec UDMA raid installation.
I plugged in the drive booted the server up, and started the rebuild in adaptecs CIO manager.
Whilst it was rebuilding, i was chatting to the receptionist when i hear a long BEEEEEEEEEEEEEP.
I head to the server room and see win2k performing a memory dump.
You know that look bunny's get when they are on the road with your headlights on them... yep... that was me.
7 hours later, I am restoring from tape backup, with yet another product that doesn't QUITE do the job right...
It'll probably be dawn before I am out of here.
Do you know exactly what the problem is? I've never had a problem with my software raid setup.
Anybody have an idea why /. would reject a submission regarding this story?
Is this truly the only Earth I can live on?
Does anyone here have any exp with ami-ide raid on iwill mobos?
----- Question authority, but not ours. Hate the man, but we're not him.
Personally, I've used RAID (both SCSI and IDE) for a number of systems - and my current workstation is a P4 1.8Ghz i845 board (made by MSI) with integrated Promise EIDE RAID.
Just a few days ago, my system started randomly freezing up - but only when doing lots of disk access on the C: drive. I've seen this behavior once before, when I first built the system; I had a defective IDE drive that was getting read/write errors. I'm pretty sure I have another drive starting to go out.
This type of behavior is disappointing to me, for a system that's supposed to reduce downtime. IDE drive errors while the drive is still mostly functional (EG. spins up ok and works, other than timing out here and there when doing writes) seem to wreak havock with IDE RAID controllers. It only flags a drive as "down" if it's completely unresponsive.
While I haven't seen a higher-end SCSI RAID array behave in this fashion (freezing the whole machine if a drive temporarily malfunctions), I've had plenty of other reliability issues with them.
For example, we had numerous Dell Poweredge servers using their older PERC II RAID cards - and the controllers all started dying off after a couple years of use. The hard drives could be perfectly fine, but if you lost the controller card - you were down until you got it replaced.
It seems like a really worthwhile RAID array would include dual redundant controllers. Otherwise, the controller is your single point of possible failure.
Most IDE RAID setups seem like a gimmick to me, more than a useful feature. People just like to say they have RAID on their home PC.
FASEST IS NOT BETER THEN CHEEPER.
I BUY CHEEP AND I AM USER OF HILLBILLY'S MAXTER 200 gigabight IDE ATa100 ANd my movies get editded fastir then a televison. glad to doughnate me time.
I have a somewhat amusing story. Just happened yesterday, as a matter of fact. I've got an integrated Highpoint Raid controller, and an Array (Raid - 0) of 2 WD Special Edition 80GB drives. I've got all my information on the drive (with key information backed up onto an additional 40GB drive -- just in case).
Anyway, I went to use my machine the other morning, after waking up. You know, check the mail, play a little music, so and so and so. Only, when I went to access the Array, where the music is, the machine promply locked. Hard. After a reboot, the Highpoint card simply couldn't detect anything past my primary drive (a 20GB drive, has the OS and backups of drivers) and wouldn't even continue or allow me to boot. Now I'm thinking "Shit. One of the drives is fried." or worse. My room and my case are not exactly the most well ventilated.
I opened it up later that night, and, get this, the cables had been knocked loose. Like, very loose. As in, not attached to the controller. It turned out that everything was fine, except that I was extremely paniced for about half an hour during the day.
BTW, the computer is at the end of what amounts to a secondary bed, under a loft. The cables, I assume, got dislodged when my SO and I were playing around there. The case got kicked a few times.
That's my story. Enjoy.
anyone played with one of these?
for my money its IDE raid all the way
---- Put Sig here:
Say I have a pc with 2 ide raid adapters that are 2 channel boards with most processing done onboard like the smallest 3ware 7xxx or 8xxx boards.
/dev/sda and /dev/sdb, mirror them in software raid ?
Can I create 2 RAID-0 arrays in hardware-raid with those boards, then since they appear as single scsi drives to Linux
Would there be an obvious performance drawback to this assuming it were possible?
I have three 40gb hard drives and am looking for a similar solution. Basicly, I want 20gb for NTFS (win2k), 20gb for ext3 (linux), and 80gb of shared fat32 (raid0) space.
I'm gonna play around with this over the next few days. If I need info, can I send you an e-mail?
Have you thought about drafting a howto?
I'd rather you do it wrong, than for me to have to do it at all.
But this is probably Adaptec's fault, since they label RAID 1+0 and RAID 0+1 opposite from standard convention.
I connect different power supply lines to each of the mirrors' halves, so that one half of each mirror is powered by one line, and the other half is supplied by another line.
If a power supply fails only partially, it usually does so on one of the peripheral power lines. With the right power supply wiring, and the 2400A set up in RAID 1+0 mode, a power supply failure will not usually result in any lost data, since it will be isolated to one half of each mirror.
Power supplies have been failing on me more often than drives have lately, even when they are used well within their rated limits.
Don't power both drives of a mirror with the same peripheral power cable!!! On many power suppplies, those separate peripheral power connector lines are on separate circuits, which means one may fail while the other doesn't. It's best to spread the chances of failure out as evenly as possible across the RAID.
Two-channel IDE RAID cannot support RAID 1+0, only RAID 0+1. Four IDE channels are necessary for RAID 1+0 to be effective, because if one drive fails in a two-channel configuration, the other drive sharing the same channel can stop working too, especially if the failing drive was the master.
Adaptec also offers open-source drivers for the 2400A, while the article neglects to mention that, and in doing so implies that only 3ware and HighPoint do.
Also, the article's table has read/write speeds of the Promise FastTrak shown backwards (133 vs 100).
Nonethless, the article's comments about the 2400A's slow rebuild time are accurate. It takes around 8 hours to rebuild my 120 GB 1+0 RAID (four 60 GB 7200 RPM drives).
And keep in mind that the 2400A is a SCSI RAID solution retrofitted onto an IDE interface -- some of the 2400A's firmware is shared with Adaptec's SCSI RAID firmware. So the 2400A is not really built or optimized for IDE from the ground up.
But if you need RAID 1+0 or RAID 5 data protection, and you have 4 inexpensive IDE drives to use, the 2400A is nice. It's twice saved me from losing any data. Don't expect blazing-fast performance, though -- just consistently good performance, very low CPU usage, and very strong reliability.
How does one hook up any other IDE device when using an IDE Hardware RAID.
If you want to stick in a cd-rom/burner or anything else how does it work, everything online assumes you already know what you are doing in this respect
nich
37 - what does it stand for really...
I was wondering what on earth you were talking about, especially since a few people had seen fit to moderate you up. Your reply had nothing to do with the parent. Because of that, it seemed like incoherent babbling to me. How amusing that it was moderated up. :-)
Need a Python, C++, Unix, Linux develop
I have been fighting with a Promise TX2000 for the past 2 days on a linux system. This article couldn't have come at a better time. All I can say is ... its useless. Unless you are a complete kernel hacker you can just about forget it. If you want to upgrade the kernel on your redhat system, be prepared to loose your array. The drivers I have found are specific to your distro and the original kernel it came with. ... the TX 2000 from promise is not technicaly "hardware" raid. From what I read the drivers are writen to do most of the raid and the card itself only has some basic raid functions... see here http://archive.lug.boulder.co.us/bymonth/2002.10/m sg00201.html
Oh and one other thing
I built a Pentium 3 gaming system from almost the ground up and decided as games seem to be having longer and longer load times, I'd get an ABIT motherboard(SA6R) with highpoint raid onboard. I matched it with 2 40gig IBM drives (the ones that don't suck) and the speedup in desktop response time even in non-gaming tasks was immediately noticable. To help offset the statistically halved reliability, I have the drives mounted in a cage with 2 case fans blowing cool air over them.
That motherboard died slowly as the capacitors blew out one by one, but it didn't kill the RAID array. Since I didn't have a full backup of the entire volume, I replaced the motherboard with another ABIT motherboard (KR7A) using the next edition of the same highpoint controller, but using an Athlon CPU this time. The RAID volume was fortunately recognized immediately, however the old windows install was unrecoverable because I never could convince windows98SE setup that I was now using a VIA chipset instead of the Intel 815. I ended up deleting the windows directory, re-installing windows, and then re-installing all my software over the old installations which saved all of my data and most of my non-registry application settings. This process was reasonably quick due to the high speed of the drives.
On the VIA chipset motherboard the RAID array is slightly slower than on the Intel board, however VIA released a driver that recovered most of the RAID speed. It's still easily the fastest responsing computer in the house out of 4 systems, primarily due to the hard drive speed.
I don't recommend RAID 0 for anyone but the hard-core hardware tweakers because the potential for rather amazing difficulties is rather high. If the motherboard dies, I will lose the array if I can't find another motherboard or controller card that recognizes the existing array format. That's easy now, but might not be easy next week or next year. Add double the statistical failure rate on top of that (remember I'm running IBM drives, ugh), and it's definately not a solution for a system that must be reliable with quick failure recovery times.
The RAID array also makes overclocking somewhat more of a gamble. I nearly doubled my disk score under PCMark2002 with a mere 3 mhz FSB overclock, however the system also became slightly unstable and I couldn't tell if it was bad memory or the drive subsystem becoming flaky at the increased speed. Some hard drives and drive controllers are notorious for being finicky about running at higher speeds, and I've read that IBM drives in particular do not tolerate PCI bus speeds much over standard.
It sure is fast though, enough so that I don't have anything but the video card overclocked. 20-30 day uptimes on a win98SE gaming rig speaks for itself. That's horrible compared to linux, but it's outstanding for a win9x gaming rig.
I haven't had the time to try Linux on this machine and since it's both my game rig and daily-use machine and the games I play (flightsims and driving sims) don't run all that well under linux, and I don't feel like dual-booting my main rig. I run it 24/7 and in 2 years I've lost one power supply, one 80mm case fan, and one stick of DDR memory in addition to the failed ABIT motherboard. Speed is great, reliable speed takes careful parts selection.
I work for a company that makes movie trailers. Due to the nature of the bussiness we need large amounts of storage and fast acess to it. For out main editing stations we use a fiber setup but for the motion graphics department we have one 400 gb Medea external system and for that departments server we have a terabyte of sotrage on a IDE RAID solution. Both of them work great, the second system is RAID 5 with hot-swapable hard drive bays. It works great and it was around 1/5 the cost of a SCSI solution and even more for a Fiber Solution.
My little Universe is cool for the people who can fit inside it (being 250 6'4" there aren't that many who can)
I have every byte of data on the main file server on a RAID array of some sort.
/var/cache and user temp space. This is stuff that can be recovered easily in the event of a drive failure. It's not backed up; it's for the squid cache, user mp3s, unpacked source trees, and the like.
/boot is normally mounted read-only, and contains a full (text-mode) system installation with all the goodies you could want to reconstruct a messed-up system, recover tape backups, etc. This is a 6-way mirror, so killing it would be extremely difficult. LILO's mirror support means that each of those 6 drives is bootable.
There are 6 drives. Each has a common prefix (/boot and swap partitions), then the rest is divided thus:
2 are striped RAID-0 for
4 are in a RAID-10 array. That is, a pair of mirrors, striped together. This is done with two Promise IDE cards, and the mirrors span controllers, so even a controller card failure can't take the system down. (This has been experimentally verified when one acted up. Reseating it seemed to fix it, but there's a spare sitting on the shelf for next time it gets out of line.) The disaster plans also involve splitting the mirror and walking out the door with half of it if we have to evacuate for some reason. Instant, up-to-the-minute backup.
Swap is over three mirrored pairs, and the kernel stripes them, so we have the equivalent of RAID-10 there, as well.
This plus ext3 has given me a very robust system.
Using software RAID lets me easily replace any broken part of the system without woprrying that the original vendor might have gone out of business or lost interest in the RAID product.
It's saved my ass a few times already, so I'm happy.
I wonder if people here have experience with random read to RAID systems. Usually I only see specs for the sustained read/write performace for the system.
We have a couple of fileservers with RAIDs attached, usually SCSI-IDE RAID systems, some SCSI-SCSI. While the sustained read is usually equally good, i.e. only depends on the host-interface, the random read under heavy load is really crappy for the RAIDs with IDE discs. And with crappy I mean ~10MB/s not 150MB/s with sustained read on the very same system. Is that expected?
The pseudo randomness of the access pattern commes from a cluster. As clusters are popular these days, and almost always produce rondom like access patterns, shouldn't people pay more attention to that?
Any experience/insight would be welcome.
Cheers, Peter
KdenLive/PIAVE - non-linear video editing
This is when you realize how scary it is that i and o are right next to eachother on the keyboard =)
I seem to remember that there is something about the NTFS file system and Windows XP that prevents doing this. However, I would like to know more about it. There are free SID changers, so that is not a problem.
I like reading the comments here, I am humble enough to know I can always learn something. But there's something I didn't see mentioned, in all these IDE RAID setups that people describe: can you have a hot spare disk? Hot spare is critical for data reliability. If you have a large RAID 5 or RAID 0+1 (not advised, always do 1+0, whenever possible), you can do the math and see how darn important it is to have the host spare.
What good it is to have a RAID 5 without a hot spare, when you can only guard against single drive failure? So, I really hope IDE RAID supports hot spare, otherwise I question the saity of mind of the admins who implement such solutions.
As for IDE vs SCSI drives, I have to say that I will always go with SCSI, as long as I am in a multuser environment where seek times are critical. Apparently (experience shows), if you put your database space on a RAID, seek times are critical for the performance of your application. In this context, I think this review/coparison would have benefitted from a real-life aplication's benchmarking, with a database hosted on the RAID.
Sigged!
Neither the article nor the comments mention the most important feature for server RAIDs: hotplug.
While quite common in SCSI RAID IDE RAIDS often miss this feature. This is why we choose 3ware 7850 controllers that support hotplug for IDE devices.
Since version 7.52 of firmware and utilities hotplug is available for HW-RAID and JBODs (JBODS only with the command line interface).
For performance reasons we use SW-RAID5 (7+1 160 GB disks on linux), which increases write performance from about 25 MB/s to >80 MB/s (bonnie), reads are about 150 MB/s SW or HW Raid.
I'm a bit irritated about the statement that "MTBF is cut in half since you have twice the exposure to drive failure."
I would have thought that the MTBF is that of the weakest link, isn't it? For example, if I had one drive with MTBF of say 10 secs, and one with 10 years, striped to a RAID 0, the MTBF is still 10 secs, and not 5 years.
Or am I missing something?
Holger, who always hated statistics!
Do you _need_ to boot from CD as a requirement? I am sure you know, but you could load from floppy or loadlin and even if loadlin doesn't work, you could install an older version (Win98?) of Windows and load it from there. In this extreme case, you could initially use the same partition you'd use for Linux for the Win98 install, and go on from there.
Not really. It's more a matter of convenience since I try every major distro every few releases.
I just put together a 1.3TB formated, Promise 8 drive, RAID 5, with SCSI 160 Backchannel in an HPUX 11 N-CLASS box. While it's speed may not be quite as fast as a SCSI setup the fact that I can run to my local computer store for an IDE disk is reassuring to me. My customers previously only had JBOD on the system which made it hard for me to sleep at night... But i couldn't get them to spring for SCSI RAID so this has worked well in my situation. YMMV
There are some things I just miss in all these tests. There has already been another IDE-RAID test around that also lacks those important information:
1) As you will notice, at the read tests, RAID0 ist a lot faster than RAID1. This is unclear to me as with RAID1 it is possible to speedup reading by splitting the read requests to 2 disks.
2) In most tests the single disk is faster than RAID5, sometimes it's even faster than RAID0 (!). It would be interesting why this happens, if this is a RAID-specific problem or if the IDE-RAID controller are just so bad. Moreover in some tests single drive ATA-100 is faster than single drive ATA-133, this is completely unclear to me and should be explained in the article.
3) RAID5 with 4 disks would be a lot more interesting thatn RAID0 with 4 disks. Nobody would probably do 4-disk RAID0 with IDE as the risk is far to high that one disk failes and data gets lost.
4) What the article lacks most is a comparison to software RAID (e.g. with Linux.) What I assume is that Linux software RAID will probably be faster than any of these hardware RAID solutions.
5) Moreover what is mostly not considered with Hardware RAID controllers is what to do if the controller fails. I use for one of my Linux servers a SCSI-Controller by DPT (now Adaptec) that works quite well but what should I do if this controller fails? Are those controllers compatible - probably not. So, if the controller fails and if I cannot get a replacement, all data will be lost. This is a BIG PLUS for software RAID.
6) Although it's quite hard to do, a comparison to a hardware SCSI-Controller would have been VERY informative.
I have three 40gb hard drives and am looking for a similar solution. Basicly, I want 20gb for NTFS (win2k), 20gb for ext3 (linux), and 80gb of shared fat32 (raid0) space. I'm gonna play around with this over the next few days. If I need info, can I send you an e-mail?
You can send me an email, but I can't guarantee I'll be able to respond right away. It's the end of the semester and I have finals coming up.
Two things you might want to keep in mind:
-From the numbers you gave, it sounds like you want to split one 40gb disk between the two OSs and set up the other two in raid0 (just because everything divides up so nicely). It's more awkward with two OSs and three disks (I have two OSs and two disks), but if you can manage to involve all three disks in the array you'll get a significant performance boost. You won't quite max out a 32 bit 33MHz pci bus, but you'll come close (at least in sustained read transfers).
-Theoretically, you should be able to format partitions of up to 2TB with fat32; however, windows 2000 and xp will only let you format partitions of up to 32GB -- well below what you need. If you end up formatting it with ntfs, be aware that linux write support for ntfs can destroy your data, so don't enable it. Also, some distros (redhat at least) don't ship with ntfs support for legal reasons. You have to hunt around on the net for an add-in rpm or compile the kernel yourself.
Have you thought about drafting a howto?
Sorry, but I don't think there are enough people that want to share win/lin software raid setups to make it worthwhile.
This link explains how to convert a single IDE disk Linux setup into a tripple IDE disk RAID5 Linux setup (while keeping your original data).
While the 10k RPM SCSI drives (eg., Seagate 10k.6) have lower access times than the 7200 RPM ATA drives (eg., WD200BB), their big advantage for servers is due to command queing. The current crop of SCSI 10k rpm drives provide more than TWICE the I/Os per second than the current crop of ATA 7200 rpm drives in storagereview.com's File/Web server benchmarks. The Maxtor Atlas 10k drive handles 271 i/o's per sec while the WD200BB can only handle 126. The higher rotational speed of the SCSI drives only accounts for a 40% improvement, not the measured 115% improvement.
So if you have a lot of simultaneous disk accesses, then SCSI will way out-perform ATA.
go here http://tech-report.com/reviews/2002q4/ideraid/inde x.x?pg=38
to see the conclusions.. you get to avoid 130 graphs.
I want to run a RAID 0 in my home computer (running linux). it's a 2800 Athlon XP with asus mobo/512MB ram. I'm currently using a 80GB Maxtor 7200. if i were to RAID a second 80GB Maxtor, what kind of speed increase would i recieve? Is it worth the extra $$?
Hi again, my e-mail is agentk_2@yahoo.com . Hope to hear from ya :-).
Weird. Anyways, my e-mail is agentk_2@yahoo.com . Hope to hear from you :-) .
Agent_Basilisk
... for the financially challenged. We use FreeBSD + RaidFrame (originally from NetBSD) for a smallish raid5 (250 GB effective storage), using 60 GB IBM ATA disks. Works like a charm. Otherwise go for www.zero-d.com -- they have scsi IDE raid cabinets. Nifty, but more expensive.
...so why complain?
so bad that I can only suggest to avoid it when possible ! Sometimes cost-considerations force one to use IDE RAID, which is the only reason for it to exist IMHO because so far I had only trouble with it.
:-).
Well, to be precise with Promise IDE RAID, that's all I can talk about. Problem is they only release binary drivers, and if you need to run a Linux distro or just a kernel that isn't supported by them you're lost.
We had to install several servers with Promise FastTrak 133 IDE RAID controllers, and we had to run Linux and VMWare with Windows 2000 Servers on it (don't ask
Problem was, those machines where dual-processor, and the normal SuSE 8.0 SMP kernel uses PAE (aka 64 GB RAM support), but VMWare doesn't support this. But Promise only provides modules for the SMP-64GB and Uniprocessor kernel. No support for Debian or SuSE 8.1 or SuSE SMP-4GB. To make a long story short, in the end we were forced to use software RAID, which is an ugly solution but one that works.
If we had used those excellent Adapted SCSI RAID adapters from the beginning we would have saved us a LOT of trouble and time, and with time=money we would probably have spent the same amount of money plus have a solid solution that we planed from the beginning.
Which files in particular can't you copy? Can you not access the data, or does the file not work when written to the target system?
--
Benjamin Coates
Is it "Americans"?
Did your IDE setup have a decent UPS and power supply? If it didn't then you may be blaming the wrong thing for the problem.
RAID is only for disk failures, not other failures. Backup is for other failures.
18GB SCSI 10K rpm drive vs 120GB ATA 7200 rpm drive.
;).
Partition 120GB drive so that you only use the fastest 18GB of it.
Now compare random access seek times. Only seeking 15% of 120GB drive
If 120GB ATA drive is too expensive. Test with an 80GB drive.
Not sure what the results will be, but it's worth trying don't you think?
Some drives would probably be better at short seeks than others (settling time etc). Don't see much info on this tho.
Using 2.(2, 4, 5).anything, and two different Abit RAID motherboards with either a Highpoint 370 or 372 chip, and either ReiserFS or Ext3, all I've had is endless hassles, data corruption, data loss, unrestorable boots, unmountable RAID arrays (both 1+0 and 5), an endless cycle of mkraid --dangerous-no-resync, mkraid --really-force, etc. etc. and of course, the obligatory restore-from-tape. Not to mention the endless VIA hassles with DMA, hd timeouts, and now, trying to restore, osst drivers dying and panicking where they once worked. ...I think the extent of it is RAID and IDE don't work reliably on Linux. I'm *completely* losing my faith in it. OSST (OnStream tape drives) are no longer supported (the old ones haven't been updated since mid 2001), and Linux in general is starting to become more hassle than it's worth for serious use.
There are valid performance and reliability reasons for using SCSI drives instead of IDE drives; the question is whether these gains are worth the cost, not whether they are there at all.
Reasons why SCSI might be worth it:
- Spin rate. Until IDE drives gain 10k and 15k spin rates, SCSI drives will always be king in multitasking and random-access situations. 3ms seek time is so much better than 10ms that you have to use it to believe the difference.
- Reliability. IDE drives have one year or at best three year warranties. SCSI drives have five year warranties. You can run modern 15k scsi drives stacked next to each other with zero additional case fans and expect to outlast your warranty. Try that with IDE.
- Hot swap. Does anyone here know of a hot-swap IDE raid solution? I think not.
- Tagged command queuing. A SCSI drive can collect multiple drive requests and reorder them to optimize the actual physical retrieval of the bits in question. IDE drives, even if the box lists this feature, have never done TCQ particularly well. This kind of thing is impossible to benchmark because its benefits only show up under heavy multitasking, not single-tasking benchmarks.
For most people, I would agree that you would be better off buying 2GB ram or two CPUs before spending money on SCSI. However, if you already have 2GB ram and two CPUs, and you still need more, then that's when you should look into high end SCSI.I failed to see much said about compatibility issues. I know, for example, promise cards may not work with computers using a Nvida chipset for video (keeps me from using them). I would like to know what compatiblity issues those other cards have.
We use a lot of Linux Software RAID for low end servers. RAID 1 for a dial up server/VPN servers/fax servers etc is fine. They have little disk IO and we mainly just want uptime.
I wouldn't bother with a controller for IDE RAID it just seems to add complications. If more speed is needed that Software IDE RAID, I'd just bite the bullet and buy a SCSI RAID controller and drives.
Something very odd about the single-drive ATA 100 results here. Anybody have a theory? To me, it looks as though the test is broken.
Have you got your LWN subscription yet?
I've been using IDE RAID for about 5 years now, simple mirroring for data integrity. Two cables aren't such a hassle ;-)
For a modest system without heavy performance requirements it's cheaper than SCSI while still offering redundancy.
But not only drive quality is an issue, there is (or has been) crap on the controller market too.
In all that time I've [luckily] only had three defective disks (on dozens of systems), but twice it wasn't of much help. That made us move to another brand of controller later, when more became available (I'm talking about what could very well be the first IDE RAID controllers, without even a windows driver or app - they were completely transparent to the OS).
In one case, a disk had slowly degraded. The controller, instead of ringing bells and flashing lights, had somehow hidden all errors from the OS, and what's worse, the error only became (very) apparent after lots of corrupt data had been mirrored from the bad disk to the good one during a defragmentation run.
Sudden death of a disk had been tested in several ways before we started using those controllers, it was also tested if running diskeeper on the array was safe, but can anyone tell me how to simulate gradually developing bad sectors?
Second painful case: drive sizes tend to grow with time, so I used a larger disk once to replace one that started showing symptoms (this time the controller did give an alert and no corrupt data got onto the other disk, yet it was the same type of controller as the first time).
No problem: mirror rebuilt, everything worked, even though the disks were of different sizes.
Only weeks later, the other drive died (the same manufacturing flaw in both? Half of those systems are still running with both original disks after 5 years, with only these three defects on all systems toghether).
After rebuilding the image, NT gave a blue screen at boot. It turned out that it simply wouldn't work if _both_ disks had a different size than the original.
Here is my experience. I have a Promise UltraTrak100 TX4, an UltraTrak RM8000, Compaq RA7000 and a FibreChannel 10 disk array. They have been used on HP-UX, Tru64 UNIX and Linux and the performance is pretty consistent across all OS's.
Speeds are sustained write speeds.
Compaq RA7000 RAID5 (6 10K disks) - 8 MB/s
UltraTrak100 TX4 RAID5 (3 7200 RPM disks) - 7 MB/s
UltraTrak RM8000 RAID5 (3 7200 RPM disks) - 18 MB/s
FibreChannel Array RAID0 (10 10K disks) - 70 MB/s (saturates the bus)
I like using an external unit as it makes it easy to move it to another server if necessary.
However, there seems to be an issue with the Promise UltraTrak100 TX4 not queuing SCSI commands. Not sure if that is just an HP-UX problem or if it happens under Linux. It seems to really slow down the HP-UX server compared to the Compaq RA7000. I'm happy with the performance of the RM8000 connected to a Linux box. The price is consideribly less than a SCSI solution.
If you want need something faster, go with SCSI. I can't justify the cost of Fibre, unless you plan on going to a SAN.
Josh
U am surprised they overlooked Duplidisk II - (www.arcoide.com) - about $300 -- Mirrors hard drives - O/S independent - no drivers because it works at the hardware level - can mix non identical drives - we have 11 of them in our office - they have several versions - our favorite has lights on the front so we can see if a drive has failed by walking through the room.
1. we had a departmental server with a Promise IDE RAID controller, RAID 5 and 4x60gb. For a max of 125 users it worked fairly well for small files, but if you started to push around 50-100mb files the difference with SCSI RAID was really noticeable. We did have a problem with just one of the disks in the array failing, and there still being data loss though. It happened 3 times. Now that server has been retired and we have a shiny new SCSI RAID.
2. A client bought 2 dell precision workstations with the promise IDE RAID controller, level 1 (striping). the perfomance gain over IDE was not that impressive (not really up to SCSI level). and both the original cards failed, but now they seem to work fine.
3. Another client bought a dell precision with the IDE RAID, level 0 (mirroring) (paranoid). of course there was no perfomance gain, but this system seems to be rock solid as a standalone workstation with critical data.
Do not recomend Dell to small business clients. We recomended our clients purchase the best support dell had to offer, however, dell would not send out a technician until someone sat down with them on the phone and pulled cards and cables. I asked our Dell rep (we do a lot of business with dell in our own offices) how much it would cost to get "I call, you come" service. THEY DO NOT OFFER IT. (home buyers get this kind of support but business clients are expected to have a technician on staff). Some of our clients have 5 employees and think that they are in the big time!
I now recomend IBM equipment to all my clients. They cost more, but if you tell them that you don't feel comfortable opening the case, they come out and do it for you! And now I don't have do Dell's job for them!
More often than not, something ends up going wrong that would/could not have occurred had they followed my advice in the first place, and then I hear about it.
Having been there and done that, the most important thing (at least in a professional environment) is to make sure you document your recommendations! You may also wish to document your specific concerns with respect to the course being followed that's against your recommendations.
Not so important with friends (hopefully), but in a professional relationship it can be crucial to be able to whip out the e-mail/printed report you sent six months ago when the client comes back all miffed, especially if you perform regular maintenance on their network (for example) and they're now blaming you for their data loss.
In the article, the author says that Raid 0 gives better performance, while Raid 1 gives better reliability. But shouldn't reading from Raid 1 have the same performance advantages as Raid 0 when reading? You can read different parts of a file from different drives, after all...
It could be worse : you could be telling me damn lies about your benchmarks.
I personally would love to hear any ide-raid stories that slashdotters might have.
This one time, at IDE-RAID camp, a girl stucka f lue up her pussy.
articles that start out poorly lose my interest quickly...'You get more data to use, but manufacturers seem less and less willing to guarantee the integrity and safety of all those extra bytes'... manufacture never guarantee integrity and safety of data. all they guarantee is replacement of a failed drive...
So far all the discussion has been about IDE raid solutions internal to a server (eg.. a card inside the computer). I would like to mention a second IDE raid solution that has worked exceptionally well for the company I work for. We started out with an SX-6000 card but found that the early version had a bug that crashed the PCI bus under very heavy network load. (email me for all the sordid details) We then moved up to the RM-8000 and SX-8000. The difference is that they, like the sx-4k and sx-6k have the i960 CPU and up to 128MB of cache, but unlike the sx4k, they have 4 seperate ide controller chips. Also, unlike the cards covered in the review, they're external chassis that attach to the server via an U160 SCSI interface. None of the OS incompatibility or driver issues apply. Drop in your favorite SCSI controller (Adaptec 66mhz U160 card in our case) and plug in the chassis. We have two configurations running here. One uses 8 Maxtor 160GB drives (Raid 6 w/hot spare, 996GB formatted space) and the other uses 120s in the same config (660GB of formatted space). We've been running these systems since March of this year in a very high load data center for both internal and customer facing applications. Performance has been splendid and even when we lost a drive, the resync time was in the neighborhood of 6-8 hours for the 1TB server. If you don't like the cards or have problems, then you might want to look into one of these boxes. Total cost (with 2 cold spares) was less than $5k in March for a terabyte. Add another 2k for the server (with gigabit copper nic) and you have a hell of a good file server for your business. Backups can be a problem with a system of this nature, however. The cheapest solution is to buy two RM/SX 8000 boxes and populate them identically. plug them both into your scsi card and mirror them. :)
Another solution is to get a DLT8000 or other large capacity tape drive and let it run for 20+ hours at a time.
(4 drive DLT8000 autoloaders will back this up in 6-7 hours but cost $50k).
Just wanted to throw a bit of fuel on the fires of discussion here.
Walking Bear
It's ok if you do actually use master and slave, but you try and work out what the feck's going on when CSEL is involved...!
I help out a broadcast non-profit group that works with tons of audio files. As one might imagine space is a very big issue with audio and since they are non-profit price is always the most important factor.
d id =556
They had to build a very large streaming server and luckily write speed wasnt an issue. Every week, at least 2 gigs of new content gets loaded onto the server.
I did TONS of research... SNAP servers were two small, NETAPPS way too expensive and SCSI drives, for some reason, are still outrageously pricey ($700 for a 73 gig drive?!?!?!?!).
I was able to make a nice RAID 5 box with over 800 gigs of available space for around $5000. I can't imagine making such a setup using SCSI for under $12k. Since the server was built awhile back, everything is cheaper now, so there's no reason why one couldnt make such a server now for close to $4,000.
Unfortunately, the server was built a few months before the 160 gig maxtors were available... the same setup with these drives would go over a terabyte.
Here's the recipe we used:
1) get one very snazzy case with lots of fans that can hold up to 16 IDE drives:
Case:
http://www.rackmountpro.com/productpage.cfm?pro
2) since price is an issue, we of course went with AMD instead of Intel:
Motherboard:
Tyan Tiger 760 MPX S2466N DDR (AMD)
CPU
2 X AMD Athlon MP 1900+
3)RAM: 2 x 512 Megs DDR from crucial
4)HD: 9 X Western Digital 120GB drives (8 for RAID 5 + 1 spare)
5)Controller: 3WARE Escalade 7850 ATA Raid Controller
When the first 8 disks fill up, all we have to do is throw in one more Escalade, and put 8 new drives into the hotswap cages and we've doubled capacity. Total cost for an additional terabyte: $1800.
The performance of this setup using Red Hat 7.3 has been AMAZING. Our server regularly streams to over 400 listeners and we're not seeing any bottlenecks on I/O. The 3ware management tools are great... should one of my ide drives go bad on me, it sends an email.
If I had to worry about a big DB with tons of real time writing going on I would think twice about using IDE but for general file serving I can't think of a good reason to justify the price of SCSI.
One caveat: as some here have pointed out here, cabling is a bit of nightmare.
I've been working on x86-based servers a long, long time.
:-)
There are many reasons one should choose SCSI over IDE, but I want to counter a few of the arguments I've read through the many messages here:
Argument #1:
SCSI can have 15 devices per bus, but why buy more smaller and more expensive SCSI drives instead of getting fewer large IDE drives?
Answer: Bigger isn't always better. On large RAID systems (real servers, here people...not Mp3 servers) one of the concepts of RAID5 is to spread out the data among as many drive spindles as possible. This keeps each drive's load level under control, and eliminates hot-spots on individual disks. If you sit down with any SAN vendor, like EMC, they will tell you the same thing.
Argument #2
Sustained IDE Raid performance can equal SCSI
This is absolutely incorrect. This may be true on a server with no CPU load. Try this again on a server running SQL and averaging 85% load. You will NOT see the same performance out of an IDE disk layer. There is simply too much CPU overhead on an IDE-based RAID system for heavy-load systems. The idea behind a SCSI controller is that it is free of the system's CPU as a bottleneck. The money saved on non-SCSI hardware will instead need to be spent on faster CPUs.
Argument #3
IDE Disks are just as reliable as SCSI
Again, completely false. You get what you pay for. SCSI disks have logic on each disk to control the operations OF that disk. In a RAID array, you want each disk to be completely independant of the others. IDE RAID requires the controller to do all the monitoring (if there is any) of each disk, lowering performance of its primary function--controlling disk I/O. Anyone who has worked on a Compaq server and used Insight Manager will be able to see the advantages of SCSI disks directly. SCSI disks will be more reliable since they are built to be more reliable. IDE disks are meant for cheap deployment on cheap systems.
Thank you, have a nice day
-brain
The article talks about OS support, and lists supported OSes. It
seems people who use less common OSes are fresh out (though FreeBSD
gets half-decent support). My question is, why does it make any
difference what OS you use? Shouldn't the array be presented to
the OS (indeed, to the BIOS) as a single disk? Isn't that the whole
point of hardware RAID (as opposed to software RAID)? Why does the
OS even need to know there _is_ a RAID? Doesn't make sense to me.
I want to be able to set up a RAID (1, 3, or 5 are the schemes I'd be
interested in), partition the resulting "drive" N ways, and install
whatever OSes I want (provided they can run on x86 and support IDE,
of course), just like I can on a single physical disk. Why won't
that work?
Cut that out, or I will ship you to Norilsk in a box.
We installed the Rainbow 100 card and all I can say is that when it goes down don't expect anything to recover.
I was all psyched when I got a mobo that supported IDE RAID. Turns out it was rather limited in what it could do, so I decided to implement it using Linux software instead. I ended up using LVM and RAID to create a structure of LVM on RAID1 on RAID0 with a spare disk hanging out in the case to boot off of. That took most of week to implement, but it's been working great ever since.
Windows XP will not copy its own registry file, for example. Without the registry, the OS is useless.
However, this behavior does not affect the ability to do a sector-by-sector partition copy under Linux. I don't know if that is possible, or what Linux tools to use.
High Point 404
We needed the biggest disk array we could get for a project without spending more than about $USD25k.
We built (3 load-balanced) machines with 4 120GB IDE/100 disks on them and used software RAID5 over them (using Promise TX2 cards).
The raw volume does about 50MB/s.
I loaded up 240GB of database onto it - this was around 200,000,000 (two hundred million) rows in a single PostgreSQL table.
I then wrote a program to randomly chose rows from the table. We could select (completely at random) 65 rows/second. I was amazed. With all the random seeking going on, the drives were still doing about 20MB/s throughput, according to iostat!
Since we needed to perform at 5 rows/second, this was wonderful!
The best price we could have done it with SCSI was about 10x as much, due to the small size of modern SCSI disks...
Although you're unlikely to read this now, I concede that you have a point. I did get a little carried away in the frenzy, as we're all prone to do sometimes, but I've seen some rather strange timelines here before, and the numbers here seemed to paint a slightly scary picture. You can sometimes learn a lot from paying attention to the small details, and I think I just climbed a bit of a learning curve.
;-)
If there were some sort of generic meta-post forum, I suppose I could have posted there - it would have been a better way to make this comment.
BTW, I still like my idea for a filter that eliminates posts made too soon after a story was posted, after all it would have caught me too.
Guess it all depends on the hardware. I've not used the Escalage, nor have I used IDE-RAID 5, or 10. I do agree with IBM SCSI drives though. They are just flat out unreliable. Have several of them in Toshiba Magna servers. I also have Quantuim, Seagate & Fujitsu SCSI drives in my HP NetServer Pro's. Both the Toshiba's & HP's use the Mylex 160 RAID controllers. Lets just say I've had 4 of 8 IBM SCSI drives go tits up in less than 3 years while all the other servers are still going strong with only 1 failure in 5 years out of 32 drives...
:) Since it's my time and THIER money SCSI...SCSI....ON SITE ON SITE :)
So maybe I should check out that Escalade card. Just for myself of course. Our company doesn't order ANY servers or equipment if it doesn't come with an on site warranty FROM the manufacturer. They aren't to up on custom built servers and such. Get into custom hardware and our software vendors act like they are working on Typhoid Mary or something. It is their money and keeps me from getting that late night/weekend call that the server is dead NOW FIX IT. Much easier to go ok, call HP and meet rep for him/her to fix it then I restore from tape. Lets time, less hassle.
All boils down to what you value, your data, your time, or both
Maybe, just maybe, 12 years from now, when HP/Compaq, or other high end server manufactures are willing to put the same warranty on the server with IDE-RAID as they do with SCSI. Then and only then I may be able to persuade purchasing to go with it. Till then it is mearly "new technology" that looks/sounds and may even be good. It just isn't a proven reliable/valuable technology.
I've nothing of importance to say, now go away before I taunt you with a second sig!
It was sad to see that 3ware is still very fast for reads, but fails to write at a reasonable speed. We had the same problems with the 6800 series and the 7400 series. Luckily they work great with software RAID. I know it sounds funny to set up a RAID card as a just a bunch of disk and do the RAID in software, but it works great. I wish this review had tried software RAID on with these cards.
I'm writing this on an MSI KT3 Ultra ARU/xP2100 with 2 x 60 GB Maxtor stripe-set. I really like the FastTrak133 controller, and the whole system boots into XP in less than 30 seconds. This is the first system I've ever had that felt fast for more than two weeks (the autobahn effect). I'm still happy with it after 5 months, by far the all time record.
Since I'm not an illiterate moron, I store all my data on mirrored and duplexed volumes on a dual cpu / ecc ram / dual power supply / ups'd / DAT backed up fileserver.
Whining about a stupid user trick and somehow blaming it on your RAID controller is pretty funny. Since it's not redundant in a stripe set, what makes you think it has anything to do with RAID? Your controller was perfectly capable of creating a mirrored set, you know.
I was the human tape changer 12 years ago on a SUN 490 with one of the very first RAID5 arrays. The purpose of RAID3/4/5 was to get the biggest volumes from the cheapest (most inexpensive) drives without sacrificing data integrity. It was never for performance reasons, and it still needed backup.
Sheesh.
I agree. Can you recommend some software to me? I'll try it.
I recently setup a two disk, 120gig 7200rpm RAID 0. Maximum speed is 87 mB/second, but the average bandwidth is an impressive 75 megabytes/second.
Faster than any single drive on the planet, even those new 200gig ATA133 drives or 15,000rpm SCSI drives.
The best part was the $200 price. Each drive was only $99 (thanksgiving sales), and the RAID controller is built-on the mb. Even if it wasn't I bought a PCI RAID controller before I bought the board for only $20 on eBay.
I haven't forgotten about backups though: I have a third 120gig 5400rpm drive to backup any important data on the other two drives on a nightly basis.
Maybe I'm mistaken, but why not just buy two IDE drives and set them up for RAID 1. If one IDE drive fails you have the second drive, and you saved yourself $800.
Hi
:(
we use Promise Fastrack with two 40gb harddrive on a RAID 1 array and we are fully satisfied.
no particular slowdown, maybe except at installation time. but i've install the OS on a single drive and then create the array.
i've not been able to install lilo on the array. i still boot on a floppy disk
(any help here will be much appreciated.)
any comment, question are welcome.
somekool no spam at mytradecenter dot com
There is third-party software that does work (mostly) for copying Windows XP sector by sector. Looks like that is better.
LOL when I first saw 7133 I thought is said 1337 ;) been use to reading too much hacker text on slashdot.
16 drives and cables and power and a dual Xeon MB in a 4U case??? Seems a little tight.
And they are already supported by FreeBSD using the twe driver. That's always a plus.
For $149.00 for the card and $79 each for 2 120 GIG, 7200 RPM, 8 MEG cache on board hard drives, that's a damn good deal.
The day Microsoft creates a product that doesn't suck, it will be known as the Microsoft Vaccuum Cleaner!
With the assistance of our student Technology intern (a HS senior) I recently built a 1TB Lunix storage server for about $5k. It uses a 3Ware IDE switch, has 7 180Gb drives for storage and an 8th drive as a hot spare.
Going SCSI for this would have required 2 additional drives for the same storage and would have added another $8000 to the price tag. This project would never have been approved if it were going to be over $13,000.
IDE-RAID systems definately have their place. At 7200k RPM, they certainly aren't as fast as 10K drives, but our application falls well within the transfer and seek specs possible with the 3Ware card.
I will soon be trying the newest version of Ghost. I've been trying PowerQuest Drive Image and DeployCenter, and it is rather flaky the way it is designed. The old version of Ghost (4 years ago) were worse.
If scientific reasoning were limited to the logical processes of
arithmetic, we should not get very far in our understanding of the physical
world. One might as well attempt to grasp the game of poker entirely by
the use of the mathematics of probability.
-- Vannevar Bush
- this post brought to you by the Automated Last Post Generator...