Slashdot Mirror
When Will Your Hard Drive Fail?
jfruh writes: Tech writer Andy Patrizio suffered his most catastrophic hard drive failure in 25 years of computing recently, which prompted him to delve into the questions of which hard drives fail and when. One intriguing theory behind some failure rates involve a crisis in the industry that arose from the massive 2011 floods in Thailand, home to the global hard drive industry.
...would be my guess.
I see even classic Slashdot is now pretty much unusable on dial up anymore.
When you least expect it, except if you expect it to happen just before taking a backup. Then it'll happen just when you expect it to.
Death, taxes, hard-drive failure.
Then it won't matter if when your drive/PC fails. Him having a backup on the same machine is almost as bad as not having one at all, IMO.
Is millions of external HDs being hastily plugged in and spinning up.
Short answer: If you actually care, you need better backups.
:)
If the HDD in one of my PCs dies, I don't care in the least. Restore it from last night's backup to the NAS, and call it good.
If up to two of the HDDs in my NAS die, I buy new ones, swap them in, resilver them, and call it good.
If my entire NAS dies, I would start to get worried, but at that point I can still fully recover (at least to where I left everything last night) from my partially offsite backup, an exact snapshot of my NAS that lives in my detached garage.
If my house and garage somehow both get destroyed at the same time, I would lose a lot, but do still have my most important data mirrored offsite... Though at that point, I probably have more important things to worry about than re-ripping my music library.
But if you care about when any one particular drive will fail on you, you've already accepted the eventual catastrophic failure and loss of your life's work as entirely acceptable.
This is months old and probably one of the first things to come up when you do a google search on hard drive failure statistics. Also the blog linked to is not the original story.
This is where the actual data comes from... https://www.backblaze.com/blog/best-hard-drive/
that writes these "I lost everything hard drive failures"? You would think people who have been in the computer industry for a decade or longer would understand the importance of backups.
Simple rules
1) Automatic. Because if it is a manual backup, it won't happen.
2) At least 2 backups
3) One copy offsite
I've had my drive up and running for over five years! This hard drive failure FUD is way overbl
If you need to ask yourself WHEN it will fail, that is the wrong question. The right question is "are you ready for imminent hard drive failure?"
If you are not running under the assumption that your hard drives will randomly fail, you have already lost. I have 20 year old drives still spinning, and 2 month old drives turned paper-weights.
Stop reading! Back that drive up!
Check how long your warranty is good for. It'll fail about a week after that.
When Will Your Hard Drive Fail?
The exact time of the next hard drive failure is about as easy to predict as an earthquake. However, there is a well know law of physics which states that the more time that passes from your last backup the more likely your hard disk is to fail more or less regardless of the dis's age and the odds of the damn thing failing increase exponentially if you have been doing something really important and/or time consuming in the interval.
2.5" Hitachi and Toshiba drives are extremely good.
It mentions on the horrendous failure rates of Seagate 3TB drives. I can personally confirm such thing, as my Seagate 3TB drive choked and started to die out on me. The drive technically still works... but it's having major issues trying to read random spots on the drive. It's not even half filled and yet the drive is running like a half-dead entity. After looking through Backblaze's articles, I noticed that the same model was used there and it had a horrible failure rate too.
What baffles me is how two big hard drive companies, Seagate and Western Digital, could produce terrible or mediocre drives after they've gobbled up Samsung's HD division and Hitachi's GST division. I hastily rushed out to buy a HGST 4TB drive (too big since I could personally live with 2TB) after seeing that I could buy one locally. So far, so good. But only time will tell if I end up on the short end of the stick again. (My fingers are crossed)
I hope your drives are smaller than 2 TB or you should be on RAID-6
Good-bye
Friends, I am hear to bear witness! Once I lived in darkness and sin, and I have suffered the pain of my transgressions.
I kept one copy--yes friends--one copy only-- of data I believed was important.
One fateful day Providence saw fit to show me the error of my ways. The foundation of rust on which my data was build collapsed into the sea of oblivion and the data thereon was lost forever to the void. Yes, Lost! Lost and without hope of salvation!
But this was a blessing, Friends, a blessing and a revelation for it was at that moment of humiliation and regret that the truth was shown to this poor sinner!
(cue rising electric organ chord)
That data is gone and despite our mournful remembrance of our departed files, they can never be brought back from their eternal sleep.
But, friends, that data was not important. For verily it is written that none may know the hour that the data will be lost, only that the data will be lost. And it is also written that data of which there is only one copy is not important data.
Brethren (and Sistren...) do not repeat my error and sin! Learn from my sin and my shame and join me in salvation!
Use ye a robust and mature filesystem with many protective features as self-checking, and multiple parity.
Yea, I say unto you multiple parity. Spend ye a small sum today for truly I say to ye that if ye are afraid to purchase an additional drive, then surely professional data recovery is beyond your means! Trust not in single parity for it is written that filesystems have grow huge in our greed for virtual machines, high resolution, and hoarding. Yea, though RAID5 was once a stalwart guardian against the failure of a single drive, RAID5 is dead and its promises are vanity for surely on the day the first drive fails thou shalt begin to rebuild thy array and before thou canst complete thy task the second drive shall fail and on that day there will be no salvation but only the wailing and gnashing of teeth.
Scrub ye regularly thy filesystem and furthermore perform regular smart long tests, but never at the same time, for it is written that a scrub and a long test shall persist unto eternity and never complete.
Implement well thy automated email notices and read thoroughly thy notices every week. When thy status report does not arrive at the appointed hour and when thy daemon sends thee an unexpected email, remain not idle but take action to investigate and resolve thy anomalies.
When thee hast constructed and filled a robust and well-monitored filesystem name it thy primary file server and do not rest in false security, but instead do the same a second time call this thy secondary file server. Locate ye thy secondary server in a place separate and apart from thy primary server and schedule ye regular backups from the primary to the backup filesystem. Monitor ye well the status of the backups and should the report of successful replication fail to arrive at the appointed time, investigate thy primary and secondary servers and all the links between. For in truth it is written that RAID is not a backup.
Go forth in peace my brothers and sisters in the knowledge that while it is inevitable that thy data will still someday be lost, this day will come to pass after all else has been lost and on this day the data will truly not be important.
"Reality is that which, when you stop believing in it, doesn't go away." - Philip K. Dick
For Mac users, time machine is a complete no-brainer. RAID won't protect you against data corruption...but time machine will. Stick it on a NAS and you'll be fine. Then use BackBlaze or CrashPlan to back that NAS up offsite. Heck, there are crashplan clients for synology systems, so there's no excuse. And it's cheap! Would you rather rebuild your whole music library from scratch, or pay $60/year for some insurance? Hello!
Note that you probably don't want to back up your TM folder.
...But I had to remember to make manual backups via drag and drop. ...
With a backup strategy like the one he describes in place, it is amazing his data have survived this long.
.
His backup strategy is worse than non-existent. It gave him a false comfort.
A second drive in the same computer? Wow, just fuckin' wow.
I just made a note to never, ever read anything else Andy Patrizio writes. It is writers such as he who give tech writers in industry magazines a bad name.
geesh.
I've had controllers failures on RAID that ended up corruptiong a lot of data despite the drives functioning propertly. So I've switched mainly to tapes. For my home use I managed to get two DAT/DDS drives for a reasonable price and bought a case of tapes, should last me quite some time. My backups are quite small. I don't need to backup my movies, since I already have the DVDs of them. For me 36GB of photos is a few life times worth. And tax pdfs, source code, and various musings tend to compress really well. But obviously these smallish tapes are not a solution for everyone.
“Common sense is not so common.” — Voltaire
TIL is a fairly common, in recent years, initialism for "today I learned." The very first result I got for a Google search for "til" is the TIL Reddit. The very second result was an urban dictionary entry. I don't know why I bothered to do the work for you; I suppose I felt generous being your Google-bot.
TIL people still don't know what TIL means...
I work for a small repair/IT firm and our experience largely matches backblazes when looking at brands. We rarely deal with NAS drives or larger capacities so I cannot speak to those. There is a 2007 study by google into predicting HDD failure. Per their data, about 50% of drives fail with a discernible warning in SMART. However, that warning requires manual watching as what they saw is that any type of pending or reallocated sector is indicative of failure. However, a few pending sectors may not be past the manufacturers threshold for failure. Therefore, 50% will fail with no warning, that is a given. The number probably approaches 95% or more if you are relying on tools that compare SMART stats with the manufacturer thresholds. You need tools that allow you to set your own thresholds or look at the numbers manually on a regular basis. My personal (and uneducated) assumption about this is that most of the pending/reallcoated sectors are caused by the magnetic domains on the surface of the platters weakening over time. Given the areal density of modern disks, slight defects in the coating or other chemical degradation could be to blame. Basically this would be a form of bit rot, and makes a sort of sense given the failure rate seems to spike for all manufactures at about the same time frame. Lastly, all drives will fail. Also other events happen, be it fire, theft, crypto viruses that encrypt your files and local backups, accidental or malicious deletion, etc. An on-site backup protects against none of those in any reliable way. Add in the fact that SSDs (which also will fail for other reasons), and are more difficult (expensive) to recover from are gaining traction, an off-site backup is the most logical solution, be it cloud, safe deposit box, etc. I am not here to advertise so I won't name names, but the solution my firm sells is cloud-based with both file and system image backups, including versioning and archiving. It also allows for a local copy of the backup set to be stored on a suitable drive. This allows for super fast recovery of large backup sets, with the online version as a backup. The backup set is fully encrypted with a choice of encryption standards and the ability to have only the customer have the encryption key (normally we keep the key as well, but we do not have to). If you are serious about your data, you should look into features like that for yourself. Relying on manual on-site backups can only be a recipe for eventual disaster.
Silence is a state of mime.
While interesting, the devil is in the details, and these are more less all generalizations. So yeah, don't go out and buy a Seagate built with those specific specs. That said, usually you can't tell wtf you are buying until you have it at home, crack the box and look at the serial numbers and such, at which point you are probably SOL anyway as it is yours now.
The analysis looks at the Hitachi drives as the best, which were acquired by WD. However they were acquired by Hitachi from IBM before that. IBM had it's own scandal for anyone that cares to remember for 1) The "DeathStar" class of drives that had an industry worst failure rate once upon a time for whatever reason, and 2) leaked documents about warranties and planned obsolesce, in that an approximate 3 year failure rate was more less built into drives for commercial reasons (i.e. to sell more hard drives). They were designed for 3 year lifespans, though they didn't intentionally fail after that.
Anyway it all depends. Certain drives, made at certain times, made by certain manufactures, *may* have higher or lower failure rates... This is why this topic is so hard to pin down...
I liken it to back in the golden age of OC CPU, people would be very particular to get lots or batches of certain CPU that would perform much better than their counterparts. However it had the same issue. You buy it, usually without knowing that kind of detail, roll the dice, and hope when you open the box it is the right serial number, etc...
Though where the similarity is really close it is by regional manufacturing. I vaguely recall some Intel CPU being make in Thailand, and others being made in Malaysia, and one being better than the other back in the day for a certain spec... I doubt it is much of a causality leap to infer that the drives made in China may be of lesser quality than those made in Thailand during that period of time...
One other thing to remember with computer electronic is binning. Usually in *any* electronics manufacturing process there will be binning where after QA testing, a product could pass, it could fail, or it could marginally pass and be classified as another product. As you may recall, after the whole Thailand flood, either for real or imaginary (for profit), there was a shortage of drives, and the prices doubled, then tripled. It would be VERY hard for any company to not cheat a bit in the binning process when the profit is triple what they used to make. So perhaps usually drives that might otherwise be binned as marginal or failiure, were making it to market simply because the drive you used to pay 70$ for is now selling for 300$ and that is too good to pass up (particularly for short term CEO getting quarterly bonuses not overly concerned with long term implications of branding).
No theres three
WD/HGST
Samsung/Seagate
Toshiba
When WD bought out hitatchi's HDD buisness (which got renamed to HGST in the process) the regulators wouldn't allow them to keep the 3.5 inch drive part of the buisness as that would reduce the number of players to two. So that part of the buisness was sold to Toshiba (who already made 2.5 inch drives). http://www.anandtech.com/show/...
I would also note that having the same corporate overlord does not nessacerally imply having the same quality or lack thereof.
Samsung owns Seagate.
You got that backwards, Seagate bought samsung's HDD buisness.
note: i'm known as plugwash most places but i screwd up registering that here somehow in the past and now can't register
I work at an MSP and we put out hundreds of 3TB Seagate drives in the field and we are experiencing this a LOT. Like, 20 - 30% failure or more within 3 years. It feels like we get a failure a day with these 3TB drives. (probably 1 every week or two)
if you are rich enough to qualify for an estate tax, then you are rich enough to afford lawyers to tell you how to avoid estate taxes.....
HA! I just wasted some of your bandwidth with a frivolous sig!
hard drives will fail whenever it will cause the largest disruption, be it money, time, career, or life safety. they got it from further development of the "critical detector" in all office copy machines, which invariably takes the machine down for days when you absolutely MUST make a squillion sets of a critical document.
if this is supposed to be a new economy, how come they still want my old fashioned money?
the hard drive crashes you!
What is Google-bot?
I'm a good cook. I'm a fantastic eater. - Steven Brust
TIL that if I want to know something, I can just ask this praxis guy who will look it up for me.
Don't waste your vote! Vote for whoever you want, unless you live in a swing state it won't matter anyways