Slashdot Mirror
For First Three Years, Consumer Hard Drives As Reliable As Enterprise Drives
nk497 writes "Consumer hard drives don't fail any more often than enterprise-grade hardware — despite the price difference. That's according to online storage firm Backblaze, which uses a mix of both types of drive. It studied its own hardware, finding consumer hard-drives had a failure rate of 4.2%, while enterprise-grade drives failed at a rate of 4.6%. CEO Gleb Budman noted: 'It turns out that the consumer drive failure rate does go up after three years, but all three of the first three years are pretty good,' he notes. 'We have no data on enterprise drives older than two years, so we don't know if they will also have an increase in failure rate. It could be that the vaunted reliability of enterprise drives kicks in after two years, but because we haven't seen any of that reliability in the first two years, I'm skeptical.'"
SSDs are all the rage now!!!!!! Who uses harddrives anymore???
At my company all the hardware is managed by CSC. They retire severs in about 3 years...including the drives.
When Fascism comes to America, it will call itself Anti-Fascism, and tell you to give up your guns.
"Enterprise" drives may have longer warranty coverage, so you are essentially just buying an extended warranty that is built into the selling price. This is how water heaters are priced...a 5 year warranty water heater is often identical to a 10 year warranty unit, but the manufacturer has crunched the failure rate numbers and will just wind up replacing a percentage of 10 year models when they start to leak in 8 years.
"We make our world significant by the courage of our questions and by the depth of our answers." Carl Sagan
What? There's absolutely difference between 87 octane and 92+ octane. While many high end cars are able to compensate for this difference by sacrificing efficiency, it's certainly not wise to put the lower grade gasoline in a high performance vehicle. Not a good analogy at all.
While I'd support an argument that the price difference between regular and premium gas is too great given its effects I'd say that arguing there is no measurable difference in their effects is beyond silly.
Depends entirely on your car.
For many cars, premium/high octane gas does very little. For higher-end cars and sports cars, it can make a huge difference.
And then on the really high-end there's a reason they make racing fuel (118 octane), because it makes a huge difference for some things.
A 1996 Buick, not so much. A Porsche or something like that, I bet it makes a huge difference -- both in performance and engine longevity.
Lost at C:>. Found at C.
What? There's absolutely difference between 87 octane and 92+ octane.
For 99% of cars, there is no difference. Unless a car is specifically designed to use a higher compression ratio, there is no benefit whatsoever to a higher octane rating. Besides, you are assuming that the premium gas actually has a higher octane rating. Years ago, it actually cost more to make high octane gas. Today the octane rating can be tweaked with cheap additives. So it is common to just make it all 92, then just use one tanker truck to make the delivery and just fill all the tanks with identical gas.
I thought this was already common knowledge. "Enterprise" drives are just a way to separate stupid people from their money. Sort of like "premium" gasoline.
No, it's already common knowledge that consumers do not and will not hammer hard drives anywhere near what commercial storage systems will, thus making this entire comparison fucking pointless when discussing failure rates.
And using data center controlled environments vs. a laptop bouncing around in a purse is hardly an accurate excuse.
enterprise drives have this thing called a warranty
i send a log to HP, and within 4 hours i have a new drive delivered to me. and they all have NCQ while most consumer drives dont
Premium gasoline is different from regular, and some cars do require it to keep working properly. That many people improperly think it's worth the price in their 15 year old Civic isn't the fault of the people selling the gas. That's like saying SSDs aren't worth the money just because some idiot stuck it in a budget system running Vista on a Pentium II.
Personally, we get enterprise grade drives at work for performance and support reasons more than reliability. As long as the RAID is configured properly, swapping out dead drives doesn't even rank "nuisance" on my list of common tasks.
Of course the more interesting quantity is miles per dollar. So if you divide those values by dollars/tank, what do you get?
Google already published many detailed reports on various issues surrounding the HDD business, proving that the money saved by buying cheaper hard-drives, and using them in data 'defending' situations (replicating data on multiple drives) made far more sense then using so-called 'enterprise' class equipment in complex, expensive configurations. Once again, to the surprise of no alpha, the KISS (keep it simple, stupid) principle wins out in engineering.
The buzz wordy, mock intellectual, synthetically complex world of 'enterprise' solutions is designed to appeal to the mind of the 'beta', a class of technocrat for whom rote-learning is everything. IT people are mostly of this class, so the 'paraphernalia' and 'jargon' make such people feel 'special'. The fundamentals of Computer Science fly right over the heads of most people involved in computer decision making.
It shames people to not even understand why the capitalist society works best with mass manufactured items, and that limited run items will always have significant compromises. Make more of an item, and it gets cheaper AND more reliable through necessity of efficiency.
But only a few days back, in some forum, people were dribbling in ecstasy because some fake enterprise HDD (RED series from Seagate?) was being 'discounted' to only 40% above the cost of the cheapest quality 3TB HDD. Many people gave EXPENSE as the primary reason for buying the vastly inferior Xbox One over the PS4 (in other words they were 'big' individuals because they could afford the more expensive console).
"a laptop bouncing around in a purse"
And I thought my fiancee's purse was huge, but even she can't fit her laptop in with enough room for it to bounce.
Perhaps you are missing this part:
Enterprise drives do have one advantage: longer warranties. That’s a benefit only if the higher price you pay for the longer warranty is less that what you expect to spend on replacing the drive.
Businesses want longer warranties especially these days as computers are being used longer and longer before being replaced. Realistically the first part to fail on a PC will be the hard drive.
Well, there's spam egg sausage and spam, that's not got much spam in it.
Premium gasoline is different from regular, and some cars do require it to keep working properly. That many people improperly think it's worth the price in their 15 year old Civic isn't the fault of the people selling the gas. That's like saying SSDs aren't worth the money just because some idiot stuck it in a budget system running Vista on a Pentium II.
Personally, we get enterprise grade drives at work for performance and support reasons more than reliability. As long as the RAID is configured properly, swapping out dead drives doesn't even rank "nuisance" on my list of common tasks.
Anyone trying to run Vista on a Pentium II, even without a "capacitor plague" bedeviled motherboard, needs all the help they can get.
I see even classic Slashdot is now pretty much unusable on dial up anymore.
> Unless a car is specifically designed to use a higher compression ratio...snip ... Which is a large number of higher end vehicles coming out of europe and aisa.
Disclaimer: Backblaze engineer here. I don't think all "commercial storage systems" get exactly the same "hammering". Some commercial systems are used to store data quietly for a long time (let's say online backup or shutterfly storage of photos), some commercial systems are hammered constantly (google's homepage search). I reject the concept that "enterprise" or "commercial" is a thing. You MUST look at the specific application. Some consumers use their hard drives quite a bit, some don't. Some corporations are hammering away at their drives, some are not.
That and to be fare, Enterprise drives may have a much higher level of usage then personal drives.
If something is so important that you feel the need to post it on the internet... It probably isn't that important.
But consumer hard drives are so much cheaper that it's not really cost effective anymore to buy Enterprise drives. You may need to replace them more often, but as SATA are hot swappable and everyone is using some variation of RAID these days, one could argue that buying Enterprise drives is an unnecessary expense. In a down economy, that might be significant.
Oliver's law of assumed responsibility: If you're seen fixing it, you will be blamed for breaking it.
Perhaps it's due to the smaller components or faster spindles creating more heat, but I rarely get a few years of service out of a single SATA drive before smartctl starts showing problems or a raid array tossing a drive. Seagate and OCZ have always been awesome about replacing the drive under warranty but still. Seems like those 400 meg IDE drives of yesteryear lasted decades before making any clicks-of-death.
Join the Slashcott! Feb 10 thru Feb 17!
Even my old 4-banger (gutless) 1997 Saturn SL1 sees a difference in pickup between 87 and 89 octane fuels when at highway speeds.
Laissez lire, et laissez danser; ces deux amusements ne feront jamais de mal au monde. - Voltaire
"Realistically the first part to fail on a PC will be the hard drive."
Only because the user isn't technically part of the PC.
I thought this was already common knowledge. "Enterprise" drives are just a way to separate stupid people from their money.
No. Enterprise drives come with:
- Longer warranties
- TLER (Time Limited Error Recovery) which means they won't drop out of RAID arrays or hang the array like a consumer drive would
- Work better in high vibration environments, like when you have 20+ drives in a single cabinet
Now, unless you need the warranty and are putting the drive into a RAID array, then you can just use the consumer level drive.
Also FYI the octane requirement can be related to timing advance, where a lower-compression turbocharged engine with more advanced timing would need higher octane gas to make longer burns from each spark (higher octane gas burns longer than lower octane gas). The earlier spark sets off a longer-burn time of gas timed to the timing, needing the longer-burn ability of the 92+ octane. An old simple truck with 0 BDC timing would be happy with 87 octane, where a newer engine with 15 BDC timing advance would be better with 92+ octane.
Fuck this is way off topic from hard drives, sorry. Just needed to fill in some missing info.
As for hard drives, the more, the better. RAID is for safety now, and SSD's are for speed where we used to have RAID-0. ETC
It is ironic that we have poor math skills displayed by Nerds on slashdot. Turbo charged cars and supercharged cars both have a higher compression ratio than naturally aspirated cars. Turbo charged cars represent more than 1% of cars sold today. QED, you are demonstrably incorrect. Actually, two of my naturally aspirated cars have compression ratios in excess of 11:1, and specifically require premium.
Premium gas does have a higher octane rating - it shows right there on the pump as well as the method used to calculate it. Can you provide support for your contention that what says 87 octane on the pump is actually 92 octane?
For some vehicles, a tuner can have a setting where one needs 92 octane... but the MPG gains are significant enough to offset the higher cost for premium.
However, this is definitely a YMMV item in the literal sense.
If drives fail at 2 percent per year, your warranty should probably not cost you more than 2 percent additional money, right? Otherwise you might as well just stock spare drives and skip the part about including HP. How much does your enterprise drive warranty cost?
from the article they are using consumer, and enterprise dives for the same purpose, so comparison is not pointless at all.
Are you saying that the enterprise drives last longer? Or just that they are replaced for free when they die at the same or higher rates? If you want to save money, I think the answer is *NOT* buy the warranty (so buy consumer drives) because the warranty costs more than just replacing the failed drives?
Perhaps you are missing this part:
Enterprise drives do have one advantage: longer warranties...
Businesses want longer warranties especially these days ...
The warranties are just more evidence that "enterprise drives" are a scam. Warranties are almost never worth the price you pay for them. If they were, few companies would be foolish enough to offer them.
All the newer shelves came preloaded with Coraid-approved drives. As I said, there's hundreds of drives involved here, a lot of SATA 1TB and 2TB and some SAS 600GB. I think out of the later drives, we've had two fail. Maybe three.
Asked about it, Coraid said, yes, the warranty is better on "Enterprise-class" or "RAID-class" drives, but also, the firmware is different. They claim that drives intended for the consumer / SOHO market spend a lot of time retrying marginal reads before declaring an unreadable sector and sparing it. They say that SAN-class drives limit the retry time, because the array controller handles it more efficiently, since it has the big-picture view.
The also say that the drives are optimized for close-quarters operation, all jammed together in an array, handling vibration and heat build-up slightly differently, and that they have minor differences to keep lubrication from migrating out of the spindle bearing under continuous operation. I don't know but I imagine loss of spindle bearing lube would add vibration and make any but the best reads more marginal.
I don't know for sure, but we've spent a great deal of US dollars on their products and our experience has borne out the fact that there's a definite difference in arrays.
As for corporate desktop and/or server use, well, I don't really know. Our servers that have one to four drives were mostly shipped with those drives, so we didn't choose them. I can't tell you if they are enterprise class drives, but I imagine they are, based on the replacement costs. And I know about what some of those costs are, or anyhow I know they were way more than I personally pay for drives for home desktop and server use. I know that because occasionally they fail, and I have to buy new ones.
Buick nerd here:
It might actually matter greatly on your 1996 Buick, as there were some available with the L67 engine, which is a 3.8L V6 that's Supercharged. Since it's forced induction, 91 or greater octane was required to reduce knock. 1996 Park Avenue Ultra and Riviera were both available with the L67.
However, your point still stands as it was probably intended: boring, generic, naturally aspirated, low compression cars rarely need high octane fuel.
You just put yourself in the stupid column. Higher octane gasoline resists detonation. The more timing you can add, the more power and fuel economy you will get. The higher the compression ratio & temperature the higher the octane requirements. Piston aircraft use 100LL and 100/130. Why? Because you don't drive at 75% throttle in your car. Race cars use 100-116 octane fuel. High compression sports cars, and anyone who's modified the spark curve via an aftermarket tuner to gain that 10-15hp require premium. Anyone pulling a trailer in mountainous terrain in the summer would benefit.
Manufacturers require 92 or better for certain high compression engines. This trend is increasing as manufacturers turn to more compression and forced induction to achieve CAFE fleet averages.
"Enterprise" grade drives are often faster, having better processors and more cache, and they don't do dumbass things like park heads every 8 seconds because the drive manufacturers have to listen to server and storage array manufacturers and meet their requirements to get certification for use in advanced storage systems.
You're an idiot. Please, stay away from any important systems. Just spend your time poasting on slashdot so you don't do any (more) damage.
once or twice i've had drives fail within a day or two of each other in the same RAID5 array. having a replacement on site FAST can be the difference between drinking beer at night or losing tens of millions of $$$ of data, spending hours restoring it and losing business in the mean time
that 6TB database i have might take 2 days to restore and in the meantime customers won't be able to access their data
There are multiple tanks within that gasoline tanker so it isn't just one single ~15,000 gallon tank. Most of those trucks have 4 tanks with 2 or 3 being filled with the 87 octane 1 filled with the premium grade and sometimes one filled with mid grade. Also I haven't seen an invoice from the refinery stating that what was delivered was all the same octane as the prices were broken down by octane. Then again I haven't worked at a gas station in 15 years so practices may have changed but if such a thing is happening they it seems like a perfect case for a state Attorney General to get involved in.
Time to offend someone
Let's assume a 10 gallon tank, and $1/gal for 87 and $1.20/gal for 92 (typical price difference in the US). It takes $10 to fill up on 87, $12 on 92. $10 / 250 miles = $0.04 per mile. $12 / 320 miles = $0.0375 per mile. So yes, there is a cost savings, though, very small per mile. You'll typically see some wear and tear (read: maintenance) savings on the engine as well, since there should be less build-up, etc. due to the higher octane.
Of course, that all assume the GP's numbers are true.
the HP drives are guaranteed to work in our HP servers with RAID controllers. no spending weeks doing money work figuring out why something doesn't work
we call in and stuff is replaced since everything is HP. no blame game saying its the other manufacturer's fault
Yes, that 80 dollar yearly difference for premium fuel is such a bank breaker. If it is, you'll never get the point because you are poor and stupid. Which leads to the chicken or the egg. Are you poor first, then stupid, or stupid first, causing poorness? Hmmm, get some researchers on this!
No difference between enterprise and home HDD's that I know of.
As for what "hammering and heavy use " of a drive is?
The biggest killer of HDD's is something called the CSS test cycle.
CSS = Contact Start Stop where the drive is booted up, spun up, and then shut down repetitively.
Generally, a HDD sitting there spinning away is not what kill them off,
however turning them on-off-on-off a lot is the most abusive thing that you can do.
I still think WD makes the best quality out there, but that's just my opinion.
just my 0.02 worth...
www.effectiveelectrons.com "chips that work" Analog, RF, Mixed Signal
While consumer drives overall might not have a significantly higher failure rate than enterprise ones, I can think of a few differences:
a) consumer drives are nearly always 7200rpm for normal models, 5400rpm for green or laptop models which directly influences the number of small random disk operations that can be performed per second and the overall maximum throughput. Enterprise drives typically range from 18000rpm at the very high end to 7200rpm at the absolutely lowest end, with 10K rpm probably the most common for bulk storage and 15000rpm for data intensive settings.
b) Enterprise drives are usually available for multiple connection types (fiber/SAS/SATA) whereas consumer drives are nearly always SATA only.
c) For some drive vendors, SMART reporting is much more consistant for enterprise drives. Also, the number of extra sectors on the drive made available for bad blocks to ensure the full capacity of the drive and to remap defective sectors can be significantly higher.
d) The newest difference between enterprise and consumer drives is that some manufactures are intentionally disabling typical enterprise firmware features on the consumer models, drive commands that are helpful for hardware raid/etc.
e) Guarenteed repair warranties on enterprise drives are frequently at least 1-2yrs longer than consumer
f) More attention is usually given to the impact of constant drive usage in the design of enterprise drives than consumer. While the average failure rate for drives in a 2-3yr timeframe may not be that different, I wouldn't be surprised if usage patterns over 5-10yrs resulted in a significant divergence. It's not that unusual for enterprise storage systems with dozens to hundreds of drives being in operation for at least 5 years, and frequently 10 years.
Also, I'd be curious about temperature variation tolerence. The longer a drive survives, the more likely it is to be exposed at least at some point to a brief period when normal building a/c fails or the computer chassis fans/etc fail....Not a few datacenter drive replacements have been required after datacenters have had power blips that resulted in a/c going offline for 10-20 minutes. This may not be a big deal for modern consumer drives where usage is relatively minimal and the drive is at least partially in low power mode.
sorry buddy. Mid grade is mixed ON SITE at the gas station. It is a mix of the regular grade and the premium. The pumps do the mixing themselves.
Don't believe you. You know that there is less energy in a gallon of higher-octane fuel than a gallon of lower-octane fuel, right? Higher efficiency and power through increased compression ratio and more advanced timing provides a net benefit to cars that require higher-octane fuel, but no advantage to cars that are not tuned to use the higher-octane fuel.
The real "Libtards" are the Libertarians!
Our Dell shelves (billing servers and store customer account info) have hot spares already spinning inside the shelves. NetApp Filers do this also. If a drive fails, the storage system begins IMMEDIATELY transitioning to the spare. So I agree with you wholeheartedly there. Backblaze uses RAID6 for the customer backup storage where we group 15 drives into a RAID group with 2 parity drives. So we can lose any 2 drives out of 15 and the data is still 100% intact. I really, REALLY cannot recommend RAID5 to anybody. Having a lone hard drive is fine for some applications (my laptop), and having RAID6 with 2 parity drives is fine for some applications. I cannot imagine why you would have RAID because you care about your uptime, but not care enough to use more than RAID5.
You're wrong. Let me tell you why.
1) There is a large difference between octane ratings. Octane is a burn inhibitor. The simplest explanation is that 87 octane gas has a lower flash-point than 91 octane gas (never seen 92 on the label at the pump, it's always 87-89-91 for the three grades of fuel at the pump). A low-compression engine is going to lag terribly when you step on the pedal with 91 octane gas in the tank because the energy output per volume is lowered by the octane inhibiting the reaction. A high-compression engine is going to knock like a SOB with 87 octane gas because the fuel is going to ignite from piston compression before the spark fires, and, more importantly, before the crankshaft has fully come around. And if you think the 4-point difference between the grades of gas can't be all that much of a difference, think again. AvGas is typically only in the 105-120 range, and it's used in turbine engines with compression levels that would reduce a car's ICE to shrapnel.
2) The octane rating can be tweaked upward by adding octane. Shocker. The cost of doing that on an industrial scale is not zero. It does cost something to add it. Removing it is an even more expensive process.
3) Delivery tankers can (and do) have multiple compartments. Just because it's one truck doesn't mean it isn't hauling multiple types of cargo. Your assertion is almost as absurd as saying the same thing about a non-liquid delivery truck. Think if every UPS truck was required to only haul one type of boxed item. Traffic would suck, to say the least.
4) The expense of having underground tanks certified by the DNR would be far simpler if it was just one tank. That isn't the case, because there are multiple grades of fuel in those tanks. Building a gas station and maintaining DNR certification for the tanks would be hella simple (in comparison to what it is now) if there was only one type of fuel.
Consumer drives have this thing called being half the price, keep one spare, what the heck if it breaks go out and buy a new one, in 1 a hour, still faster than 4 hours. What kind of enterprise organization wouldn't have a few hard drives spare just in case a few failed. Send the old one back to replaced, in their own good time.
I don't see why you would have to pay 100% markup for what is basically insurance, for the manufactures defects.
Sort of like airline tickets that you can reschedule, more than 2x the price and still subject to availability (last time my company bought one), just buy the non refundable ticket, if your plans change then buy another one, the average cost is going to be less, unless you change your plans a lot, perhaps you need better planning? You also have travel insurance for such things which is not the cost of the plane ticket, and covers other things too.
No, from TFA:
, so the comparison is indeed pointless (more accurately, it's baseless).
make imaginary.friends COUNT=100 VISIBLE=false
Are you saying that the enterprise drives last longer?
I didn't say that.
Or just that they are replaced for free when they die at the same or higher rates? If you want to save money, I think the answer is *NOT* buy the warranty (so buy consumer drives) because the warranty costs more than just replacing the failed drives?
If your company wants to do that, then do it. But I would think that is a hard sell to the IT directors who want service and replacement parts quickly. Here's the scenario:
1. HD fails
2. Log ticket with HD company and get replacement drive with little cost
or
2. Put in a purchase order for a new drive.
At some companies, buying a new drive outright is more troublesome/bureaucratic than getting a replacement drive.
Well, there's spam egg sausage and spam, that's not got much spam in it.
No it doesn't. You utterly fail to understand what the octane rating means. The engine in your saturn would in no way benefit from the higher octane rating. It could in fact run without noticing a problem with a significantly lower octane rating. Octane ratings matter in high compression engines or turbo/supercharged engines, not in econobox.
Persistent Volume manager for Kubernetes - https://github.com/dwimsey/openshift-pvmanager
Sometime those 15 year old civics do benefit from premium if they developed hot spots in the chambers from neglect/abuse. Also refiners like to have the higher grades having higher additive content so running a tankful or premium through them periodically probably cleans out some of the deposits in the engine so it runs more like it did new.
Time to offend someone
RAID5 array
losing tens of millions of $$$ of data, spending hours restoring
6TB database
I'd try replacing whoever set this up with someone competent.
I hope not... You do know "purse" is slang for a hooker's honey hole, right?
Do you even lift?
These aren't the 'roids you're looking for.
It is a common misconception (pushed by the drive manufacturers) that RAID arrays need Enterprise drives. RAID stands for "Redundant Array of ***INEXPENSIVE** Drives". The whole idea is you write a software layer that deals with the failures and limitations of the cheap drives. If your RAID software cannot handle independent drive failures, precisely what value is that layer adding? We have not seen "Enterprise" drives work better in a high vibration environment. It's an old joke but worth repeating: You know how you can tell if a hard drive salesperson/company representative is lying to you? Watch their lips closely, if their lips are moving, they are lying.
I wonder how many more slashdot stories will be based upon the same Backblaze story of the "first of its kind" (ignoring Google's older paper) story on hard drive longevity, that doesn't name names?
Disclaimer: Backblaze engineer here. I don't think all "commercial storage systems" get exactly the same "hammering". Some commercial systems are used to store data quietly for a long time (let's say online backup or shutterfly storage of photos), some commercial systems are hammered constantly (google's homepage search). I reject the concept that "enterprise" or "commercial" is a thing. You MUST look at the specific application. Some consumers use their hard drives quite a bit, some don't. Some corporations are hammering away at their drives, some are not.
Why is this not +5 already? He is exactly right in that all workloads do not fit neatly into the containers the marketing people seem to think they do.
> AvGas is typically only in the 105-120 range, and it's used in turbine engines with compression levels that would reduce a car's ICE to shrapnel.
No. Standard avgas is 100LL (100 Octane, low lead). It's just like gasoline from the auto pumps, but the octane level is higher and there's still some lead in it, where the auto industry is fully unleaded. (There are some aircraft that can handle lower-octane mogas, but they're rare.)
Turbine engines typically use Jet-A, which is a diesel-like, kerosene based fuel.
The warranties are just more evidence that "enterprise drives" are a scam. Warranties are almost never worth the price you pay for them. If they were, few companies would be foolish enough to offer them.
Translation: I personally have never benefitted from a warranty so they are useless to everyone.
At the many IT departments I have worked, warranties were used extensively. From little things like memory to whole motherboards were replaced without hassle. The only major company I know that uses consumer grade HDs in volume is probably Google and that is only because they have designed their server infrastructure to use massively identical and disposable hardware.
Well, there's spam egg sausage and spam, that's not got much spam in it.
"Enterprise" grade drives are often faster, having better processors and more cache
The cache is whatever is written on the drive, so a "Enterprise" drive with 32 MB of cache has less than a "Consumer" drive with 64 MB. I don't know what the heck you think the word "Enterprise" gets you in this case?
drive manufacturers have to listen to server and storage array manufacturers and meet their requirements
Different storage arrays have different requirements, I hate the idea that people think "Enterprise" magically got all the tradeoffs correct. For example, low power and high responsiveness are BOTH valid goals but probably are at odds. Some Enterprises (like Backblaze and Shutterfly) care deeply about their electrical power bill and the drives aren't the performance bottleneck. Should we buy enterprise drives or not?
Source on the tanker claim?
Also FYI the octane requirement can be related to timing advance, where a lower-compression turbocharged engine with more advanced timing would need higher octane gas to make longer burns from each spark (higher octane gas burns longer than lower octane gas). The earlier spark sets off a longer-burn time of gas timed to the timing, needing the longer-burn ability of the 92+ octane. An old simple truck with 0 BDC timing would be happy with 87 octane, where a newer engine with 15 BDC timing advance would be better with 92+ octane.
While you are correct your numbers are off.. i haven't seen a car thats less than 10 years old with timing at 0 BDC or retard.. a naturally aspirated 90's miata runs 36 advance, and you can safely take that to 39 advance on 93oct and into the low 40's with 100oct
'...if only "Jumping to a Conclusion" was an event in the Olympics.'
I hope not... You do know "purse" is slang for a hooker's honey hole, right?
Puckering one's lips is slang for a rug maker's beehive?
I think you've got your slang backwards -- puckered lips are puckered lips no matter what kind of lips they are. That's not slang, it's descriptive. A handbag "purse" is borrowing the other definition to describe a bag with a puckered opening -- and has since slid into use with a variety of similar clasp-fastened handbags.
Back to the GP's observation... very few pursed handbags are large enough to hold a full-size laptop.
The drive companies claim that the enterprise drives are designed to work at a higher level of usage. However, at Backblaze we have been running both the 25,000 consumer hard drives and the enterprise drives in this study 24x7.
Let's assume a 10 gallon tank, and $1/gal for 87 and $1.20/gal for 92 (typical price difference in the US). It takes $10 to fill up on 87, $12 on 92. $10 / 250 miles = $0.04 per mile. $12 / 320 miles = $0.0375 per mile. So yes, there is a cost savings, though, very small per mile. You'll typically see some wear and tear (read: maintenance) savings on the engine as well, since there should be less build-up, etc. due to the higher octane.
Of course, that all assume the GP's numbers are true.
Well we know that your numbers are not. Good luck finding premium gas at $.20 higher than regular. It's usually more like $.30 - $.40 higher, at least.
"Somebody has to do something. It's just incredibly pathetic it has to be us."
--- Jerry Garcia
Yes, many of the vendors require you to use "their" hard drives. For example, at Backblaze, the Dell storage systems we use for for the central servers have "Dell" drives. Realistically, those are simply WD or Seagate drives with a different badge. Regardless, the failure rate of these drives that are "by Dell, for Dell, in a Dell"...still fail more often that plain 'ol consumer drives.
Orig GP here :).
They are true for *this* car I own. It was on the order of about 40 miles on the previous car. Each car is different as it depends on a bunch of factors. Also this car starts up right away with the high octane. Takes a few cranks of the starter with the lower stuff. It is tuned for the higher octane.
For me it is more about time. Less time spent at a gas station. I am willing to pay more for that.
Funny side affect is it also lets you figure out which gas stations are really selling cheap stuff as high grade.
Also gasbuddy is your friend :)
I totally agree that "bureaucracy affects IT decisions". In a previous company we sold spam blocking software (we were the good guys) but our customers asked us to provide the software and hardware in a bundle because they had a hard time convincing their management to purchase stand alone computer hardware. So we pre-bought a PC clone, marked it up by a FACTOR OF 4 (for our trouble), put a sticker on the front with our company name and the IT guys happily passed the price on to their managers who happily signed the P.O.
Here in Australia, 92 is the standard fuel and 97 is the premium. I can't imagine putting 87 in my car...
All correct.
Everything else being equal, higher compression engines have higher efficiency and higher compression requires higher octane gas. There is an efficiency / cost tradeoff with compression and reuqired octane that resulted in 87 being most common, but 92 octane being used in cars, and 100 octane being standard for almost all aircraft piston engines.
Turbines and diesels use similar (often identical) fuels that are completely different from gasoline. Diesels are high compression, but rely on the fuel spontaneously igniting when it is injected, as opposed to gas engines that rely on the fuel not igniting util the spark fires. Turbines are fairly low compression and relatively low efficiency, but much better power to weight which makes them better for high power aircraft use.
Diesels in general are more efficient than spark ignition engines, but (all else being equal) are heavier and more expensive.
For tons of information read the Taylor, "the internal combustion engine in theory and practice" .
At my gas station the mid-grade is also 10% Ethanol. I don't think you can get there just by mixing regular and premium...
Any sect, cult, or religion will legislate its creed into law if it acquires the political power to do so.
The only major company I know that uses consumer grade HDs in volume is probably Google
What qualifies as "major"? :-) This article is about Backblaze, we have 25,000 consumer hard drives, are we "major"?
Given the cheap PSU's I've seen in a lot of boxes (and the rate of failure), I'd say in many cases that it's a contest between the drives and the PSU, especially when you get to areas with flakey power.
Agree 100% with you. And if you have 100 drives you can buy a mere 5 spares, which temporarily costs you 5% extra vs. the 100% extra that the enterprise drives cost up-front...and both are likely to fail at the same rate.
Exactly - companies make money on "extended warranties"; that's why they offer them (and why many companies push them so hard.) Like insurance, they are doing math to determine a price point where the average failures cost less than the amount they charge. If a single failure is not catastrophic to you (for example, your house burning down), you shouldn't pay for the risk-adjusted extra cost.
Exactly - they come with all those things. However, at Backblaze we run both our 25,000 consumer hard drives and the enterprise drives in RAID arrays, and the consumer drives are run in enclosures that have MORE vibrations => and still the consumer drives performed better from a reliability perspective.
I think you missed his point. With the money you save, buy a spare drive.
6 drives with enterprise warranty: $1800, 12 hour replacement
7 drives with consumer warranty: $1300, instant replacement
You can't use a consumer drive in a RAID array if that drive will spend 90 seconds trying to recover a normal read error before sparing the sector out. TLER means "give up almost immediately" on media errors.
Yes, it's a bit of a scam that you have to buy a high-end drive to get TLER, since it's just a flag in the firmware, but it's still critical ro RAID.
Socialism: a lie told by totalitarians and believed by fools.
Besides that premium gasoline can be required for some high end cars, "enterprise" drives usually have SAS connectors which are required in a lot of environments (eg. multi-host systems).
Also most enterprise drives have different performance characteristics - for example if a drive read or write fails on a sector, a desktop drive will time out and retry for 10-20 times or more, an enterprise drive will return after 2 or 3 times and report it as unreadable which can make a great difference in performance (waiting for the drive to return with a failure state in 200ms vs 20ms). Usually multiple sectors are damaged and requested so this can easily make the difference between your SAN repeatedly stalling for 1 or 2 seconds at a time or not.
Custom electronics and digital signage for your business: www.evcircuits.com
Concerned about cost?!? This will be the same dipshit buying a single $7 pack of cigarettes, and a $2 bottle of Coke, telling you that the cost of $20 a year for premium is too much.
You can't use a consumer drive in a RAID array if that drive will spend 90 seconds trying to recover a normal read error before sparing the sector out. TLER means "give up almost immediately" on media errors.
That's only for crummy RAID controllers with no memory to speak of. Use ZFS with simple host adapters - you'll be happier and save a bunch of money.
SAS IDENTIFY is the only useful feature missing from SATA. Well, full duplex too, but you better be buying $$$ SSD's if you have those concerns, so SAS vs. SATA shouldn't matter that much.
My God, it's Full of Source!
OUTSIDE_IP=$(dig +short my.ip @outsideip.net)
Any car with a modern ECU will advance timing until the knock sensor detects detonation. With higher octane gas, detonation occurs at higher thresholds, so the timing can be advanced a few degrees farther than it could with 87 octane. So yeah, even an old Saturn 4-banger can get better performance out of higher octane.
Consumer drives have this thing called being half the price, keep one spare, what the heck if it breaks go out and buy a new one, in 1 a hour, still faster than 4 hours. What kind of enterprise organization wouldn't have a few hard drives spare just in case a few failed.
The Fortune-sized one I used to work at, alas. And the RAID drives would always blow out while I was on vacation. Then before I got back, another RAID drive would blow, breaking the array. And in-house inventory wouldn't have any spares and the drive in question was no longer available from approved suppliers.
We weren't just paying for the 7-year warranty, however. We were also paying for the high-performance SCSI interfaces. These systems were doing mainframe-grade work. We even had mainframe tape readers.
I was curious, so I did some math:
320 miles / 15 gallons = 21.33 MPG
250 miles / 15 gallons = 16.67 MPG
320 - 250 = a 70 mile difference in performance.
At 16.67 MPG, 70 miles equates to about 4.2 extra gallons needed to reach 320 miles. So for that person, using premium is like having an extra 4.2 gallons in his tank.
In my state, the best prices I could find for 87 and 92 gas were:
$2.83 for 87 and $3.11 for 92
$2.83 * 19.2 = $54.34
$3.11 * 15.0 = $46.65
So for every 320 miles he drives, he is basically saving $7.69. Not earth shattering, but definitely a win.
That's an unfair scenario. There could be many quality differences between enterprise and consumer drives that simply don't come up in their environment. I know when I make consumer and enterprise-grade objects, of course the consumer-grade objects work -- I don't build carp -- but the enterprise-grade work better. For many values of better. Most often, that better includes things like a wider temperature range, dirtier air, and more frequent and rougher shipping. Even my packaging is wildly different as a result. Better foam, larger boxes. Also interestingly stupid things like additional electrical certifications. And then there are emergency situations like easier repair, in this case data-rescue would be a major feature, as would fire and flood resistance..
the HP drives are guaranteed to work in our HP servers with RAID controllers. no spending weeks doing money work figuring out why something doesn't work
we call in and stuff is replaced since everything is HP. no blame game saying its the other manufacturer's fault
That hardly helps when HP sends out a new critical alert for their hard drive firmware every damn month "or risk data corruption". Like I'm going to (or can) pull drives and update firmware at the drop of a hat. I went with samsung and seagate drives instead, for 1/5 the price HP wanted, and bought a couple spares instead. No firmware updates required (HD204UI's aside, which I updated prior to installing back in the day).
I don't need someone to blame, I need my stuff up and running so no one asks me why it's offline.
The blog post states: "You might object to these numbers because the usage of the drives is different. The enterprise drives are used heavily. The consumer drives are in continual use storing users’ updated files and they are up and running all the time, but the usage is lighter. "
That invalidates the conclusion they're drawing. You can't put two different types of drives under different workloads and then conclude they fail at the same rate. The fact that other studies have reached similar conclusions (Google published one a few years back) is irrelevant when it comes to evaluating whether or not *this* study has measured what it seeks to measure.
Consumer drives and enterprise drives may fail at equal rates, but using different workloads doesn't help us reach that conclusion.
Doesn't work like that. Knock sensor outputs are used to retard timing. Maximum advance is determined in advance for the engine with recommended fuel. Otherwise a failed knock sensor could quickly result in a destroyed engine.
Perhaps if OP's engine was suffering from excessive knock due to a fault (excessive carbon build-up, incorrect timing, etc.) then higher octane fuel would make a difference.
The real "Libtards" are the Libertarians!
You don't need SAS drives for TLER, just the Enterprise-grade SATA. TLER is just the correct behavior for RAID - if there's a media error on one drive, the right thing to do is immediately read the redundant copy from another drive, and re-write the problem block at the RAID controller level.. 90 seconds of retry is the right thing to do when that's your only copy of the data.
The SAS/SATA divide in the disks themselves is mostly about IOPS - SAS drives seem targeted for the IOPS-bound market (we don't care how big the drive is, we only care about IOPS per spindle), but SSD is killing that off. SAS RAID is just corporate momentum at this point, as Enterprise-grade SSD (low capacity, high re-write count) is now better across the board than SAS.
Socialism: a lie told by totalitarians and believed by fools.
That and to be fare
It's illegal to be a fare. That's slavery. Or cannibalism...
It all matters what you value - reliability or performance. EITHER ONE is valid for companies, you can't say every "Enterprise" wants drives that error faster and successfully get the data back less often. Backblaze is a company, we value reliability way way WAAAAAAAY over performance. We want the hard drive to take 90 seconds and give us the data - heck, take a full 3 days to get the data back, we'll wait, so will our customers. We have no performance problems at all - customers are extremely happy getting a successful restore FedEx'ed to them in 48 hours (one of the restore options is a $189 3 TByte hard drive sent to you anywhere in the world where you keep the hard drive).
If you think SSDs fail because a part "fails" you lack understanding of how they work.
SSDs have a property called "write endurance" - their data cells are rated to a specific number of writes. Every time you write, you consume some of the remaining write capacity of the drive. It works like a salt shaker: works find until you run out of salt.
Enterprise drives can have dozens to hundreds of times the write endurance of a consumer drive. For example, the Intel SSDs we use are rated to withstand 100% of the drive's capacity in writes every 24 hours for many years on end. A consumer drive couldn't do that for more than a few weeks, perhaps a month or two.
I'd happily pay 2x or 3x the money to get 20x the write endurance.
I have no problem with your religion until you decide it's reason to deprive others of the truth.
Until we see some names, those studies are useless...
I've got better things to do tonight than die.
Backblaze happens to use software RAID6 - standard Debian Linux, we use the built in mdadm tool. Our current pods have 8 GBytes of RAM, so I guess they could theoretically use all of that (and swap) instead of using "crummy RAID controllers with no memory to speak of".
Concerned about cost?!? This will be the same dipshit buying a single $7 pack of cigarettes, and a $2 bottle of Coke, telling you that the cost of $20 a year for premium is too much.
... says the mathematically challenged.
12,000 miles / 30 MPG (generous) * $.40 premium = $160 / year
That's fine if you're driving a sensitive car that requires or is designed for high octane gas. Otherwise, it's just being an idiot for paying more for something that you don't need.
"Somebody has to do something. It's just incredibly pathetic it has to be us."
--- Jerry Garcia
the HP drives are guaranteed to work in our HP servers with RAID controllers. no spending weeks doing money work figuring out why something doesn't work we call in and stuff is replaced since everything is HP. no blame game saying its the other manufacturer's fault
HP makes it hard to use commodity drives because they want to make a $300 markup on each unit... not because SAS is somehow lacking in standardization.
TLER is just the correct behavior for RAID - if there's a media error on one drive, the right thing to do is immediately read the redundant copy from another drive, and re-write the problem block at the RAID controller level.
Your RAID doesn't issue simultaneous reads and return whichever one gets in faster? ZFS waits forever - if the driver layer wants to fail a block read at n seconds, it's good with that. Fortunately we can tune our OS's to meet whatever requirements we need.
My God, it's Full of Source!
OUTSIDE_IP=$(dig +short my.ip @outsideip.net)
No, the point is that you get your redundancy from RAID, instead of from the drive trying it's best to pass that track under that head and hope something coherent emerges. Trying to read a damaged sector is less reliable than reading the undamaged redundant copy. It's great that you "value reliability" and all, but it doesn't sound like you've thought it through.
At any scale the only interesting question is "where are your multiple copies written - those servers aren't all in the same data center, right". Reliability of an individual server is just a cost analysis on replacement costs vs higher-end components, with the inevitable trend of "high end -> commodity -> disposable" over time. Data reliability is properly a measure of the size of the disaster you can survive, and "damage to one sector" is as uninteresting as it gets.
Socialism: a lie told by totalitarians and believed by fools.
Plus, you don't have to worry about the replacement drive coming in from a different batch that has 100 fewer sectors on it. You can tell a new RAID admin because he will accept the system's default of "use every sector on the drive" instead of reserving .1% for drive manufacturer rounding shenanigans.
I don't know, maybe the situation is better today, but if I'm put in charge of a RAID array, I'm always going to shave off a few sectors out of distrust of drive manufacturers.
I read the internet for the articles.
In the business, this is called duty cycle, and there are significant MTBF differences between enterprise quality drives (FC, SAS) and consumer drivers (SATA, NL-SAS) at high duty cycles.
If you are running a server like a SQL database, an Exchange server, or a Host like ESX or HyperV with a good number of guest vms, then the disks are going to be active much of the time. This high duty cycle will wear down the low end drives over time.
Someone you trust is one of us.
In my experience the first part to fail is usually one of the cheapass fans that are stuffed in most PCs. Most people don't notice though because there's still enough cooling in the box to keep it from dying outright. Of course the hot spot in the box leads to more component failures (especially hard drives, they hate being hot).
I read the internet for the articles.
Enterprise drives typically range from 18000rpm at the very high end...10K rpm probably the most common for bulk storage
Backblaze pays something like $45,000 / month in our electrical bill. We vastly prefer "green" drives that spin slower and use less electricity. There are many, many "Enterprise" applications in the world that are not bottle necked on spindle speed (like backup and Shutterfly-type big-data-rarely-accessed), those enterprises deserve slower drives. I guess I object to using the word "Enterprise" to describe "Fast" - why not just mark your drive as 15,000 RPM or 7,200 RPM and be done with it? No need to add the pointless label "Enterprise Drive".
SMART reporting is much more consistent for enterprise drives
No way. All hard drives do SMART reporting. Sometimes the "bridge" between the processor and the hard drives won't pass the information, so a cheap USB enclosure might be hiding the hard drive SMART stuff from you, but that isn't the hard drive's fault. In fact, we have an expensive Dell drive shelf with an LSI (?) controller that hides our enterprise drive SMART stats from us, very annoying. There is no correlation between "Enterprise" and "SMART reporting".
some manufactures are intentionally disabling typical enterprise firmware features on the consumer models, drive commands that are helpful for hardware raid
The whole concept of RAID is that it is a software layer on top of all the cheap drives. RAID doesn't require any interesting instructions. Pretty much needs to write data to an individual drive and read it back later.
I wouldn't be surprised if usage patterns over 5-10yrs resulted in a significant divergence.
Time will prove you right or wrong, we plan on updating and releasing these numbers every few years. Stay tuned....
Depends on how the RAID controllers are configured - what you describe gives better single-IO latency at the cost of fewer IOPS overall, but favoring IOPS is the norm for hardware RAID controllers. While instinctively I want to say "however Oracle does it, do it the opposite way", I can certainly understand that the former use case is important to some (at which point, yeah, no real reason to care about TLER). Sometimes you optimize for latency, sometimes for bandwidth.
Socialism: a lie told by totalitarians and believed by fools.
You can always pour some additive through your tank once in a blue moon. They are just light distillates and a bit of detergent.
Faster! Faster! Faster would be better!
Yeah, no kidding. Back in my younger and less persuasive days, we were on a project where we were forced by PHBs to use consumer drives in an enterprise system (storing and retreiving syslog data in a VERY busy environment). We were literally blowing them out every three months or so until the Powers That Be finally relented and let us put in proper storage (back then that also meant shelling out for a pricy SCSI HBA). I think that the gap has closed somewhat since then, and there are also some interesting options in drives that are purpose-built for things like DVRs and low-volume RAID. Also, back then (I don't know if it's still the case today) enterprise HDDs were tested individually for quality control, whereas consumer HDDs were just randomly sampled from each batch.
For many enterprise applications, though, the difference in things like seek times and sustained data transfer rate can be substantial in a busy environment.
Help save the critically endangered Blue Iguana
I'd happily pay 2x or 3x the money to get 20x the write endurance.
That only makes sense if you are hitting the write limits. If the drive dies because the bearings wear out after 5 years of spinning regardless of the number of writes, you have just paid 3x the money and gotten exactly zero benefit.
You don't need "Enterprise-grade" SATA either.
ATA device, with non-removable media
Model Number: TOSHIBA DT01ACA300
# smartctl -l scterc /dev/sdn
smartctl 6.2 2013-04-20 r3812 [x86_64-linux-3.11-2-amd64] (local build)
Copyright (C) 2002-13, Bruce Allen, Christian Franke, www.smartmontools.org
SCT Error Recovery Control:
Read: 70 (7.0 seconds)
Write: 70 (7.0 seconds)
Here in Australia, 92 is the standard fuel and 97 is the premium. I can't imagine putting 87 in my car...
Australia displays the "Research Octane Number" on the pumps, while the US diplays the "Anti-Knock Index", which is:
((Research Octane Number) + (Motor Octane Number)) / 2
Since MON is often 8-10 points lower for the same fuel, this results in 4-5 points lower on the pump display in the US.
Trying to read a damaged sector is less reliable than reading the undamaged redundant copy.
You're thinking about it wrong. You always want the maximum amount of information from every drive, you can choose to use that information however you like, I don't want "Enterprise" drives that won't try hard to get every last bit.
Here is an example: We have had problems reassembling / resyncing RAID arrays because one stubborn drive pops out and fails too easily (we run two parity drives - so if you are already down 2 drives a 3rd stubborn drive is a bummer). If the drive would just stay in and try harder, we could get through that particular operation. Backblaze then adds it's own end-to-end SHA-1 on every file - trust us, we'll absolutely know for certain whether or not we recovered the file accurately or not from that particular RAID array or not. But until we reassemble the RAID array and get the file system back online, we can't even check what we are holding. Fighting with it costs us IT time. Again-> we have no performance problems at all. I know this is hard for some organizations to grasp when you never seem to have enough IOPS. But the nature of online backup is not like the nature of your billing or account info database.
For example, the Intel SSDs we use are rated to withstand 100% of the drive's capacity in writes every 24 hours for many years on end. A consumer drive couldn't do that for more than a few weeks, perhaps a month or two.
Every drive tested here has handled over 1000x capacity of writes, which is around 3 years at your benchmark of 100% capacity per day. Every one of those drives is a consumer drive. Some are showing signs of eventually failing, but none has lost one byte of data.
In real-world usage, that means consumer SSDs are easily good for 10 years before write endurance becomes a problem.
You'll be glad to hear that for 250GB and larger drives HD manufacturers agreed on standardized sector counts.
Also, welcome to the 2000s.
I always assumed that Enterprise HDDs were the same as consumer ones, they just got binned differently after passing through tighter QC scrutiny or more arduous burn-in testing or something.
Either way, it's not that surprising... I mean, a drive either fails within the first several months, or runs fine until something inside's well past its MTBF. Right?
So within the first three years, we're basically talking about the near end of the curve. It's not until after 3+ years that normal HDDs start to really show their age, in my experience.
Friend: "The NIC is misconfigured..." Me: "No prob, I'll just telnet in and fix it." *Silence*
We have Dell's 4 hour support. Not only will they ship you a drive in 4 hours if you want they will send out an techie with it and they will replace it for you - great if it is a remote site. They ask for a dump of the firmware log from the array and look for any issues developing on the drives. If they spot any evidence of failing drives, punctured stripes, etc. they will ship you replacement pre-emptively. I got sent three disks when one failed recently so I could rebuild the array before any more failures occured. Fantastic service. Disks are *known* to work with RAID cards so you don't have issues with incompatible firmware. No need to dick around with enabling TLR (or equivalent) to get them to play nicely with RAID. NCQ implementations Just Work.
Aside from that if you're using SAS or Fibrechannel disks then then isn't a "consumer" alternative anyway.
I used to work for a cheap-ass company that used consumer drives in "servers". Never again.
If you're running a Hadoop cluster (or similar) then some regular disks are probably fine, but if you're running a business critical array on consumer level SATA then you're a sucker.
The number of hard drives you have does not make you a major company - what is your turnover?
Christ, dude, you're dangerously wrong and I have no idea how you got modded up. There certainly IS a difference -- run high octaine gasoline in a (normal) low compression engine and you'll ruin your exhaust valves, because higher octane burns slower and your valves will get burned. For the same reason, run regular in your high compression engine and unless you detune it, it will ruin the pistons from dieseling ("spark knock").
Use the correct gasoline for your engine, or risk expensive repairs.
You're even more wrong about the price of the extra octane. Lead is dirt cheap (or was before they found out how hazardous it is). If what they're using now was cheaper they'd have used it instead of lead; lead was used because it's cheap.. It costs a lot more to raise octane with the newer chemicals.
Free Martian Whores!
That's happened to me recently, with multi-TB drives. The last time I bought drives for an existing RAID array I checked the sector count and found that the the drive I was looking at buying had MORE sectors than the drives already in the array. That wasn't a problem, but it shows that "equal" sized drives aren't necessarily equal. Do you have a citation or some good search terms to find out about this supposed standardization? I'd love to read anything that might give some information about which manufacturers, if any, have actually standardized.
It's been said that some drives (enterprise drives?) have more spare sectors reserved. Given that that number of sectors per platter is a physical attribute they can't readily adjust, it seems that reserving more sectors would leave fewer visible sectors.
I manage a couple of petabytes worth of disks (consumer, not enterprise) for the HPC center at Vanderbilt University, and they get absolutely hammered by CMS-HI users 24/7/365. At scale, you will daily see problems that you would never even think of.
The firmware on consumer hard drives is often crap. Very few of them support TLER, we have ~400's drives (Seagates) that needed a firmware fix to prevent sudden death but the fix wouldn't work en bulk over the SAS controller so we had to yank/flash/replace/repeat, and drives will occasionally lock up hard and require a power-cycle.
Don't believe for a second that Linux doesn't need a defrag utility. We were mystified by a sudden influx of permanent drive *slot* failures. After *much* investigation, it turns out that our users were filling them 100% full, erasing 5%, refilling, erasing 5%, etc, until the average file (~100 MB) had thousands of extents. The vibration from the head frantically scanning the disk to read the file was enough to cause the SATA connector to destroy the connector on the backplane (Supermicro chassis, would *NOT* buy again, Chenbro is the way...) We wrote a simple defrag script that simply copied the worst files to a different location and then move them back.
RAID5 isn't nearly sufficient at this point because you will eventually have two or more simultaneous failures just due to the number of disks. We wrote our own filesystem to offer Reed-Solomon-6+3 redundancy.
I'd love to know if you guys have any similar "WTH" horror stories.
Miles-per-dollar divided by dollars-per-tank?
That would give tank-miles per square dollar.
Not a particularly useful metric, I would guess...
"Of course the more interesting quantity is miles per dollar. So if you divide those values by dollars/tank, what do you get?"
MULTIPLYING those quantities would give miles per tank.
But DIVIDING those quantities gives a meaningless value in tank-miles per dollar squared.
We have had problems reassembling / resyncing RAID arrays because one stubborn drive pops out and fails too easily (we run two parity drives - so if you are already down 2 drives a 3rd stubborn drive is a bummer). If the drive would just stay in and try harder,
IMO that drive should get shredded at the first sign of trouble (ideally, you'd be getting SMART alerts pre-fail at least some of the time). But of course it's very common for multiple drives in an array to fail within a short time window, due to shared environmental problems. I'd call 2 drives failure in a short window a signal that it's time to replace the server and re-sync its data to the other copies.
But until we reassemble the RAID array and get the file system back online, we can't even check what we are holding
Except for the synchronous copy on the other server(s), right? And the async off-site copies for any data that you've had for a little while?
Socialism: a lie told by totalitarians and believed by fools.
Only because the user isn't technically part of the PC.
Also the user isn't under warranty.
(that might explain because most of them are crap)
Stupid planned obsolescence. We should complain to the manufacturer.
"Sufficiently advanced satire is indistinguishable from reality." - [Tips: 1DrYakQDKCQ6y52z6QbnkxHXAocMZJE61o ]
Turbo charged cars and supercharged cars both have a higher compression ratio than naturally aspirated cars.
Wrong; compression ratio is a function of the geometry of the piston/combustion chamber only. In practice, it is actually usually the opposite - a similar engine that is turbo/super-charged will have a lower compression ratio than its NA counterpart, specifically because the forced induction requires less compression ratio to get the same power.
Even my old 4-banger (gutless) 1997 Saturn SL1 sees a difference in pickup between 87 and 89 octane fuels when at highway speeds.
You are deluding yourself. Unless your car uses high or variable compression (it doesn't), there is no benefit whatsoever to higher octane gasoline.
In the business, this is called duty cycle, and there are significant MTBF differences between enterprise quality drives (FC, SAS) and consumer drivers (SATA, NL-SAS) at high duty cycles.
No there isn't. This article was just more confirmation that there is NO difference between "enterprise" and "consumer" other than the price. Plenty of other people have looked at the data and reached the same conclusion. Also: the MTBF number on the side of the box has no connection to reality. It is just a number made up by the marketing department.
Are consumer hard drives just as reliable for the first 3 years when installed in rack servers?
Are enterprise drives more reliable past 3 years when installed in poorly-maintained desktop computers?
What the standard sizes *should* be:
250GB - 488397168 sectors
320GB - 625142448 sectors
500GB - 976773168 sectors
750GB - 1465149168 sectors
1TB - 1953525168 sectors
1.5TB - 2930277168 sectors
2TB - 3907029168 sectors
3TB - 5860533168 sectors
Don't have the figure for 4TB, neither do I have anything about 2.5" drives.
If you have a drive that appears smaller than that, check Host Protected Area and your controller settings.
If you have a drive that appears larger than that... please post the model.
For eMLC, there's also a bit of "cheating" going on.
Rated erase cycles: 20nm MLC 3000 - 20nm eMLC 15000
Min retention time after rated erase cycles: 20nm MLC 180 days - 20nm eMLC 30 days
So "enterprise" MLC flash is allowed to be run way beyond the point where "consumer" flash would already be considered too unreliable.
I haven't seen regular for $2.xx since the early 2000s. The last few years, I haven't seen it under $3.xx more than once and that was only for a day or two.
"It is a denial of justice not to stretch out a helping hand to the fallen; that is the common right of humanity."
You're wrong. There are still stations with midgrade tanks and fuel terminals still blend midgrade for loading.
once or twice i've had drives fail within a day or two of each other in the same RAID5 array. having a replacement on site FAST can be the difference between drinking beer at night or losing tens of millions of $$$ of data, spending hours restoring it and losing business in the mean time
that 6TB database i have might take 2 days to restore and in the meantime customers won't be able to access their data
True, but any IT department that doesn't already have those replaements on site ready to go is foolish, foolish, foolish. You don't call up and get a replacement drive when it fails (well, yes, you do, but not right away). First you pull a replacement from the replacements bin and you install that and then get everything happy again. Then you worry about calling about the replacement. If the replacement takes 4 hours or 8 hours or 2 days, it doesn't matter, because you've got more replacements on hand. If you plan well, that is.
Only because the user isn't technically part of the PC.
You clearly don't see the situation from the management perspective.
What the heck? The error retry and sector sparing are within the drive itself. ZFS doesn't even see this. What ZFS can see is a drive not responding for 90 seconds after a write command, and ZFS or the driver below the ZFS level does not like this. There is real danger of multiple drives being kicked out of the storage pool quickly and the whole pool failing, when proper drive behavior lets the pool continue undegraded even in the face of bad sectoirs on multiple disks.
There are plenty of consumer drives that can be set to the same TLER (time limited error control) behavior as enterprise drives, though.
Right on the money about using ZFS, though. I will never understand losers using old fashioned expensive caching RAID controllers when ZFS on dumb SATA/SAS ports is far superior in every way. Many or most of them are Windows losers, of course.
, so the comparison is indeed pointless (more accurately, it's baseless).
It's not totally pointless, but the claims made don't seem to be warranted.
It doesn't pass as a scientific study: and relying on the supposed result could be hazardous, both to your data, and to your job.
If Enterprise drives aren't any better: then why is Backblaze using them? Do they plan to retire all Enterprise and nearline drives from all manufacturers, and switch to consumer drivers, based on the results of their limited dataset?
Perhaps they would like to do a comparative study of the effects on robustness, reliability, and performance; of their storage systems, disk drives, and applications -- during and after completing that complete media rotation out of Enterprise disk drives; including how and if disk longevity, failure rates, etc, seem to change on an Array-by-Array basis.
Now... THAT if done carefully, could provide some meaningful data
Did you consider; that the Dell drives may be self-failing due to special firmware?
Whereas; the same drive mechanics in a software RAID, without the special array controller and drive firmware --- would still be "failing", but the failure could be undetected longer.
In other words --- a portion of the consumer level drives are failing, but due to the absence of array scrubbing, and special drive firmware the failure, and potentially many bad sectors exist but have gone undetected, and a fair portion of the consumer drives may be ecking out silent data corruption, unreliable performance, or worse.
In some cases; the drive appears to be working, but the undetected failure could be picked up by a full read/write test.
These 2 factors combine to kill modern drives, IMHO. The increased density makes the drive work harder to combat vibration.
HD bays and mounts ought to come with more carefully engineered dampening.
The major difference in "enterprise" NL-SAS drives versus consumer SATA drives is the native SAS command set making the drives more efficient and more compatible with RAID controllers and technologies like storage pools in Server 2012. Step up to true 10K and 15K SAS drives and you get significantly higher IOPS, slower than SSD but less expensive per GB.
This if purely anecdotal but: I've reliably found that putting 89 octane in my generic '99 Ford Ranger (which I imagine is about as far as one could get from the vehicle that 'high end car' brings to the mind's eye) gets roughly another MPG versus 87. I haven't run any sort of ANOVA or crap, but it's consistent enough to notice.
Virtualization is perhaps the biggest driver in failure rates of enterprise drives. When you have several VM's competing for access on the same spindle, you're bound to have a lot more drive wear than an HDD in an laptop running not much more than a web browser.
"... but also, the firmware is different. They claim that drives intended for the consumer / SOHO market spend a lot of time retrying marginal reads before declaring an unreadable sector and sparing it. They say that SAN-class drives limit the retry time, because the array controller handles it more efficiently, since it has the big-picture view."
What you are describing is known as TLER or "Time Limited Error Recovery" (the Western Digital name for it, at least). See TLER
For 99% of cars, there is no difference. Unless a car is specifically designed to use a higher compression ratio, there is no benefit whatsoever to a higher octane rating.
False. Putting aside differences in additives between the premium and lower grade fuels such as cleaning agents and oxygenates, premium fuel has a different calorific value to standard unleaded leading to greater fuel economy, all other conditions staying the same. This isn't intentional, it's more of a side effect of the inability to get higher octane ratings without sacrificing trading off some of the specifications which can be met with lower quality fuels. I.e. if you blend in more Alkylate and Reformate to get your higher octane rating you're likely to well exceed the HCV and RVP specs as there's no other high octane products you can blend in which have lower RVP and can bring you closer to the spec. This is called "give-away" in the oil industry.
And putting all that aside your idea that a car will experience no difference assumes a perfectly working car as per factory spec designed for that fuel. A lot of cars even with lower compression ratios will start suffering when their engine is wearing out leading to poor combustion and light knocking in which case Premium fuels offer an even greater advantage.
Besides, you are assuming that the premium gas actually has a higher octane rating. Years ago, it actually cost more to make high octane gas. Today the octane rating can be tweaked with cheap additives. So it is common to just make it all 92, then just use one tanker truck to make the delivery and just fill all the tanks with identical gas.
Could not be more False. It is illegal in most countries to measure fuels after additives are added. A premium fuel must stand on it's own ground and pass the specs, THEN permitted additives may be added to certain fuels. But then maybe you and I aren't classing "additives" the same. I.e. Tetraethyl-Lead is listed as a (former) "additive" on wikipedia. So is "tolulene". Neither of these are "additives", they are just blend stocks that gets mixed in fuel, and that still makes your argument invalid since they are expensive -products of oil refining and anything but cheap.
Disclosure: I work in a refinery.
Also I haven't seen an invoice from the refinery stating that what was delivered was all the same octane
That's because the GP doesn't have a clue. Fuel needs to meet specs for sale before they leave the refinery. In some cases additives are added at the terminal but in those cases they need to be sent off for the lab again before they are legally allowed to be sold as a fuel. Bottom line is if it's on a tanker to a retail site, it meets all the required and very stringent specs for that fuel.
Custody transfer is an incredibly strict process.
I'd happily pay 2x or 3x the money to get 20x the write endurance.
That only makes sense if you are hitting the write limits. If the drive dies because the bearings wear out after 5 years of spinning regardless of the number of writes, you have just paid 3x the money and gotten exactly zero benefit.
The poster to whom you responded said "SSDs have a property called "write endurance" - their data cells are rated to a specific number of writes. Every time you write, you consume some of the remaining write capacity of the drive. It works like a salt shaker: works find until you run out of salt.", which suggests he's talking about SSDs here, not HDDs. SSDs don't have spinning parts (well, other than the electrons, protons, and neutrons of which they're made :-)), and don't have bearings.
Fuck this is way off topic from hard drives, sorry. Just needed to fill in some missing info.
Dude, there's no such thing as off-topic on /. Remotely interesting is perfectly sufficient.
10^14 on consumer drives vs 10^15 on enterprise drives.
Worst. Signature. Ever.
Won't it wear down any drive over time?
Either way, Google uses consumer grade drives. If there is a difference, it cannot be significant enough to justify the higher cost.
$2.96 on my way to work this morning.
Cheap storage VM.
There really is a difference. However... that difference is in the firmware, and that's where the manufacturers were scamming.
There is a thing called TLER - time to recover from errors. That is, if the drive is trying to write or read from a block, and it finds a problem, it goes to write to another block, or uses error recovery to read. For servers, they really want the time that the drive keeps trying to be under seven seconds. The consumer drives could be adjusted using software like hdparm.
Then around '09, and apparently starting with WD, they made a change to the firmware, and you could no longer change that variable. Servers scream and gag and give up, and tell you the drive's dying, when instead of spending under seven seconds, the drive keeps trying for ->over two minutes-. Everything you read says do *NOT* use those for RAID, either.
The server grade drives are very much *not* into spinning down, and they have that short TLER.
This is why, around here, we're ecstatic at the new WD Red drives, that are "targeted towards NAS"; the reality is that they've got TLER set to seven seconds. And, where the server grade drives are two to three times the price of consumer grade drives, or higher (some sources are a *lot* higher), the Reds are 1.33% of consumer grade drives.
Reliability: we have some of everything, and have not noticed a real difference in reliability. And our drives get used a *lot*.
mark
The other thing to look for in enterprise SSDs is the user of "super-caps" (large capacitors), these provide enough charge that the SSD can power down gracefully if the power is cut. Such as flushing any buffers, making sure the sector mapping table is updated, or whatever other metadata they need to track.
Wolde you bothe eate your cake, and have your cake?
It makes a helluva difference when my truck is hauling a 14,000 pound load in hilly country.
~REZ~ #43301. Who'd fake being me anyway?
A safe rule of thumb that I've used over the years is to only allocate 99% of the drive size when putting together the array. This is a bit easier with Software RAID and mdadm under Linux.
You might even be safe at 99.5% or 99.9%, but the latter is probably a bit risky.
Wolde you bothe eate your cake, and have your cake?
We're switching over to putting the VMs on top of enterprise SSDs for just that reason. Even 15k SAS drives don't have the IOPS to deal with the demands of lots of small VMs that all assume they are running on dedicated hardware.
Price difference between enterprise SSD and 15k SAS drives is about 2.0-2.5x $/GB. So it's more expensive, but not an order-of-magnitude more expensive. And the gap keeps narrowing.
Wolde you bothe eate your cake, and have your cake?
I'm not sure what you mean by "turnover"? If you are asking how many customers we have, I apologize but I'm not allowed to release that number (not my fault, I would post it on our homepage with a live number if they let me!)
:-)
But I was mostly joking, I think by "major" UnknowingFool meant the largest 4 or 5 companies on earth like Google, Facebook, Apple, Microsoft and maybe Yahoo. I assure you that Backblaze is in no danger of displacing any of the members of that list.
very common for multiple drives in an array to fail within a short time window, due to shared environmental problems
Exactly. We had one interesting incident where in the middle of the night, 3 pods right next to each other in a rack all went berserk and all their RAID fell apart. That's 135 drives all at once (3 pods each with 45 hard drives). We reassembled them all, and the VERY NEXT NIGHT at the same time it happened again. We moved all three servers to different ends of the datacenter -> and finally figured out which server was causing the problems. The fan bearings on a fan were going bad, and when the fan came on it vibrated the entire cabinet. We have "nightly cleanup" jobs that run to verify data integrity and delete files we no longer want, this was enough load to cause the CPU to heat up enough to trigger the bad fan.
Yes, but there's no difference between Cadillac parts and Chevy parts, except the price.
What this is saying is that the use portion of the bathtub curve is the same between two differently marketed products produced from the same manufacturing lots/lines. In other word, "Well duh!! What else did you expect?!??".
Do tell you thought there were two different manufacturing processes: one for consumer, one for business?!?
The difference between marketed types is what tested out with sampling as more reliable or not. But that's AFTER the use portion of the bath tub curve. Infant failures are all undesirable so all are removed by burn-in anyway.
IOW not even news.
And yet over multiple trips with my Saturn SL1 between Vancouver and Seattle with a loaded car (4 passengers), I have found much better acceleration responsiveness to the application of gas at highway speeds with 89 octane vs. 87. Now maybe it's the car computer that's monitoring and interpreting the octane level as a signal that fuel economy is less of a concern and changing how it runs the engine in response to the gas pedal. But the behaviour was sufficiently noticeable and repeatable that while I use 87 octane during regular city driving, I will fill up with 89 octane on those occasions when I'm going on a long highway trip.
Laissez lire, et laissez danser; ces deux amusements ne feront jamais de mal au monde. - Voltaire
The other thing that can happen in old engines is you'll get carbon buildups which can increase pressure in the cylinder and reduce your fuel economy, not to mention pinging and knocking. The 'easy' solution (as opposed to tearing the engine apart) is to just use higher octane gas which will make that problem go away, at least for a while. However, the higher octane gas can also cause the carbon to build up faster, so while it can alleviate the side-effects, it can also make the root problem worse in the long run. But for an old beater car that may not matter.
From personal experience: Running 87 Octane my 1984 13B RX-7 got 20 MPG and had great responsiveness when needed. Running 92, 92+ octane it lagged. It also only got 17 MPG... Maybe tuning could have been changed to bring it back up... but why incur additional costs for no benefit. Unless you have an abnormally high compression ratio that requires high octane fuel, 'premium' is a waste of money.
I noticed this while the car was still less than 10 years old (closer to 5 actually). And even though it's now close to 17 years old, its mileage is very low for the car's age and it's been well maintained over that time. It's far from a beater.
Laissez lire, et laissez danser; ces deux amusements ne feront jamais de mal au monde. - Voltaire
Well, I did try it when they were launched, 10+ years ago here in Europe.
The difference was, my ordinary VW car ran longer with higher octane, and in my actual test, over months, on the same average daily commutes, things was about equilibrate: go refill less often but paying a bit more each time -at the time of my test that was indeed equivalent.
I kept on with lower octane (refilling slightly more often) because I feared high-octane prices would raise faster. But I'm not even sure this happened, indeed.
Maybe I need another test...
Herve S.