Yeah, that’s what RAID 0 does. Your data is striped across two drives so two drives are fetching half of the file at the same time, so disk operations are faster.
Back when a big drive was 2 gig, the drive could only push one or two megabits per second through the ATA interface, and the access times were slower, it was a big thing. Now, drives are so fast that the can stuff gigabits per second out through a SATA cable, and nobody even cares what the access time is since it’s so small, RAID 0 is not nearly so important.
I know that. I was just making the point that raid isn’t a scam or con.
On a seperate note: Just so I know, why don’t you guys consider raid 1 to be a backup solution? (albeit a non-standard one)
Surely if one drive fails, your backup is the other drive.
If one drive fails, you have the other one, if the controller fails, you’re screwed.
For my webserver down at the ISP, I’ve got one drive on the motherboard, and another on an external USB box. I mount the drive, back up the data, and unmount the drive. That way the controller isn’t the weak point. (Granted, if the ISP catches fire, I’m screwed, but it’s not a High value asset anyway, and there’s too much data to push over the wire offsite).
Because you can lose your computer to theft, or to fire, you might get a virus that corrupts your data (this would affect both copies), you might delete files by mistake, and so on. RAID allows quick recovery from hardware failure, but it’s not a backup.
Are you speaking of RAID for home use, or for business/enterprise use? Because if you are speaking of RAID for home use, I can at least half-see your argument. If you are saying RAID is a scam in the enterprise - that is just so far off base that I don’t know where to begin countering it.
I’m not sure I buy the RAID is a scam/not a backup argument - but what I’m wondering is why the heck isn’t tape backup more widely available as a backup media in sizes similar to what drives run. To back up my home server I’d have to run a frickin automated tape array just to back up a 3tb of data? I don’t consider Blu-ray disks usable for backups for drives that large.
Seems like tape tech (or similar rewritable media) is about the same generation as it was when I was backing up mainframes in the early 1990’s.
The research & development for tape technology hasn’t been keeping up in density of packed bits as much as hard drives.
Today, the LTO-4 tape is 800gb and a tape drive to use is $3500. Therefore, you’d have to backup more than 34 terabytes for it to become cost effective. If you’re under the 34 terabyte threshold, it’s cheaper to just buy extra hard drives and use them to backup the primary data. I gave up on tape a long time ago.
Maybe I shouldn’t have used the words “scam” and “con”. But what it was sold to us as was not what we got.
They told us it would make our drives safer, more reliable, and we wouldn’t lose important data in the event of a single drive crashing. What they didn’t mention was there was an increased likelihood of a drive crashing, and when it was offline waiting for a replacement, though your files are safe, they are inaccessible. Plus, you have less hard drive space to store things on anyway.
At least, I think that’s how it works. I don’t know, as every time I’ve encountered RAID it has completely failed to live up to any useful potential whatsoever.
It was a huge mess, and every time someone suggests it as a good idea, I warn them to stay away.
There isn’t an increased likelihood of drives crashing/failing in raid configuration. No risk increase over the risk of losing a drive in a single drive configuration.
Depending on the raid configuration being used, a raid array can absolutely be fault tolerant when a single drive in the array goes down. And the data is usually completely accessible. When the replacement drive arrives, it can be added back into the array (or a software raid rebuild done) to restore complete protection.
I think you’ve just had a bad experience with raid. Maybe the claims for raid were overstated and/or it was poorly administered.
As mentioned above, RAID 0 is of limited true benefit in many circumstances. If you need something to pump data for a high demand app…SSD’s rock at this.
I have often wondered what kinda of data throughput you can truly get from striped SSD’s (visualizing consumer grade motherboards blistering and scorching)
RAID-1 or better is a godsend in the world of small business. I have had machines with RAID-1’s fail in unbootable states but popped right back to life with a fresh drive and a rebuild of the mirror.
Striped SSDs would be pretty much irrelevant. SSDs are already designed to take advantage of the principle behind striping - the units can read multiple areas of their memory at the same time, and so the file system is set up to distribute files across the memory in such a way that multiple reading mechanisms can each read part of the requested data at the same time to increase the total rate of access.
Maybe not inherent in its make up, but both times we used it, the drives failed within weeks of setting it up.
That’s what I thought first time, but then it happened a second time, years later (after I had hoped a lot of the bugs had been ironed out) and not only did it fail in the same kind of way, when we looked it up we found a whole bunch of complaints online similarly frustrated with RAID’s limitations and inexplicable failures.
The people here who say RAID works for them are the same kind of comments that convinced us to try it both times, so forgive me if I don’t trust your word any more now than I ought to have back then.
Guanolad, you only had a problem because you used RAID incorrectly. RAID means Redundant Array of Inexpensive Disks. Note the word inexpensive, there. Hard disks are cheap.
To address your concerns:
“increased likelihood of a drive crashing” : not so. The chance of any particular drive crashing is not affected by whether it’s in a RAID configuration or not.
“when it was offline waiting for a replacement, though your files are safe, they are inaccessible” : RAID is designed for use in enterprises. It is expected that they will have spare disks on hand. Recovering from RAID disk failure should only take a few minutes, as long as you don’t have to wait for a replacement disk.
“Plus, you have less hard drive space to store things on anyway.” : Not a problem. You can use as many disks as you need, and create a huge amount of space. And the more disks you use, the larger the amount of usable space. If you use three disks, then you have 2 disks worth of usable space, and one for fault tolerance. But if you use 20 disks, then you have 19 disks worth of usable space, and 1 for fault tolerance. The loss of space is not significant.
Integrated RAID on some motherboards is A: a PITA for home users to set up and B: often very limited in scope and functionality compared to its higher quality big brothers.
Despite your experiences, RAID is, and will continue to be an invaluable tool. Shit happens and maybe you are one of the unlucky ones.
I would be very curious if you truly had drives failing faster, RAID does not do anything to amplify the failure rate its just a part that makes two drives write the same thing at the same time.
Statistically in a mirrored raid you are twice as likely to have a single hard drive because there are two drives, but you are now looking at less than half the chance of losing any data.
RAID is not about the cost of the drives, its about the value of the data. If you have an 8TB NAS unit a 500GB drive dies in it every month but has never lost any data, you still have an array that is doing its job. It could be housing information that would cost hundreds of thousands if not millions to rebuild/replace. Replacing a $70 hard drive every so often is pocket change.
If a raid makes it more likely that ANYTHING will survive a hardware failure its usually a damn good thing, even at consumer levels.
I had a customer who lost a book she was writing (two years of work) along with
several years of geaneology research, about 5000 photographs to a hard drive failure.
It was all housed on a $399 HP special. When we discussed her repair options I mentioned RAID and an offline backup solution. She was astounded that anyone would pay for such things when you could just buy another computer rather than spend another $200 on a system that would have allowed her data a soft landing.
IME many, many people have fallen into the “good computers cost about $400” mindset, and proceed to dump their data onto bargain basement PC’s.
I cannot comment on the specifics of the Office San, I handle security, they handle data storage. But they have a Rack nearly fully populated with disks. 140 and 300 Gb a piece, Probably more than 100 drives. The array has lost 4 or 5 disks this year. When a disk fails, the SAN contacts HP, they open a trouble ticket, and send us a disk.
The failed disk is safe to return to them as it contains no sensitive information that can be recreated.
This SAN contains almost all of our databases, and 140 virtual machines. In the heat of the day, it serves 2000 IOPS (I/O operations per second)
And it’s a small example, when compared to the Googles of the world, only about 10 TB (yeah, my math doesn’t add up. If you want specifics, I’ll ask the guy in the morning.)