The hard drive failure thread

[goontron]

Re: The hard drive failure thread

I have NetBSD 7.1 installed on an Eee PC 1000 (8GB SSD). The "system" (kernel, system/userland binaries, X server/clients, compiler, man pages, log files) consumes ~650MiB. The sources (kernel + all system/userland apps, S, compiler, all "supplemental packages" like OO/KDE/asterisk, etc.) consume another ~2GiB. No swap space configured (cuz it's an SSD).

I suspect this is a tiny fraction of what a BARE W7 installation occupies!

Roughly on par for Windows 7 Embedded.

[Curious.George]

Re: The hard drive failure thread

On Linux runing on a hard drive, I am using Mint on the lap top I am typeing this from . The iso is like 1.9 gig , 2 something running with all the swap files.

I have NetBSD 7.1 installed on an Eee PC 1000 (8GB SSD). The "system" (kernel, system/userland binaries, X server/clients, compiler, man pages, log files) consumes ~650MiB. The sources (kernel + all system/userland apps, S, compiler, all "supplemental packages" like OO/KDE/asterisk, etc.) consume another ~2GiB. No swap space configured (cuz it's an SSD).

I suspect this is a tiny fraction of what a BARE W7 installation occupies!

[desert-rat]

Re: The hard drive failure thread

On Linux runing on a hard drive, I am using Mint on the lap top I am typeing this from . The iso is like 1.9 gig , 2 something running with all the swap files.

[Curious.George]

Re: The hard drive failure thread

Well, you typically aren't going to need that HDD full of, say, Linux ISOs that you haven't used since 2002...

Are you SURE? Sure enough to delete them and possibly NEVER recover them, again?

Please locate a copy of the sources (or binaries!) for Mach MK78 for me. I'll wait...

...

...

No luck, eh?

I can install FreeBSD on a Compaq Portable 386 (a few MEGAbytes of RAM) -- but not if I've discarded the ISOs of the older versions of FreeBSD that DID run on hardware of that era!

When you treat disk space as "free", there is no upside to discarding things and plenty of opportunity for "subsequent regrets".

[Agent24]

Re: The hard drive failure thread

The problem is that you (typically) don't have good "predictive powers" as to what you MIGHT need at some future date.

Well, you typically aren't going to need that HDD full of, say, Linux ISOs that you haven't used since 2002...

[Curious.George]

Re: The hard drive failure thread

My backup archive is not that big and quite messy. I really should do this in a more organized fashion... but I don't care enough, I suppose. I just know that I have my most important stuff backed up fairly well (at least 1 off-line copy somewhere.), which is mostly my music, pictures, some OS ISOs, and some software and drivers that are hard to find.

The problem is that you (typically) don't have good "predictive powers" as to what you MIGHT need at some future date.

Just yesterday, I realized that a piece of software that I wrote 30+ years ago would PERFECTLY fit the bill for a project I started! (Of course, my memory of having written it probably influenced my realization that it would be a viable "solution").

So, today I have to figure out what I would have called it "back then" if I have any hope of finding it in my index/catalog. I am only belatedly realizing the value of maintaining timestamps on files! I've been careless, over the years, so have many files that I know are decades old but bear "recent" timestamps owing to how/when/where I copied them from their originals.

<frown> Just tiny pieces of "lost data" that, in this case, would have made my search much easier!! (search for year = 1986)

One thing, for sure, is that all of the files will likely have 8.3 names! :>

[momaka]

Re: The hard drive failure thread

well my wd caviar black 2tb kicked the bucket today. was trying to test a video card i bought on ebay so i booted up the system and while installing the video driver, the system was extremely sluggish copying and installing the files. i thought the sata cable was focking with me again and producing lots of ultra dma crc errors so i fired up crystal disk info to check. the cable was fine. instead i saw "caution" and it had one pending sector in smart, so it looks like the drive is failing.

I find that UDMA CRC errors often have to do either with bad cable or bad contacts between PCB and head amp / actuator. So definitely remove the PCB on the drive and check its head amp contacts for corrosion.

This may not be the problem but I have seen a lot of drives where the contacts between the board and the drive become tarnished and may be causing issues, cleaning them with an erasser my help.

Yup.
Mainly older and some newer Western Digital's had/have that problem (as always, I'm not in-line with the latest tech out there, so can't speak about current HDDs from them).
I now not only clean the contact pads with eraser, but also apply leaded solder over them.

"Only back up the data that you care about..."

+1

My backup archive is not that big and quite messy. I really should do this in a more organized fashion... but I don't care enough, I suppose. I just know that I have my most important stuff backed up fairly well (at least 1 off-line copy somewhere.), which is mostly my music, pictures, some OS ISOs, and some software and drivers that are hard to find. Everything else (game ISOs, movies, and sentimental/fun junk)... if I have backed up, it's because I probably had a convenient option at the time. If I didn't, then it's probably not backed up.

And if it all gets lost.... well F--- it, nothing lives forever, right.

[Curious.George]

Re: The hard drive failure thread

i think its more than that. its a waste of time and energy as well. i have to spend time setting up the "drive recovery and rescue environment" and i have to wait for it to pull all the data off the drive before i can use my system again.

And you never really know if you actually have recovered the data or just SOME data that appears to have satisfied the ECC and other aspects of the filesystem format. (I keep MD5's of all my files in a database isolated from the files themselves so I can verify their integrity -- online OR in a recovery attempt).

its also a waste of electrical energy as i have to leave my system on overnight so it can try to rescue and pull and reread all the data off the bad sectors. in the end, its how much time and energy u can expend to recover as much data as possible versus the loss of use of functionality of the machine u want to use while it performs the recovery.

And, unless the platters/head assembly physically seize up, you're never quite sure when to "give up" in those efforts. Will another hour of tinkering yield more data? Or, will it just be an hour wasted? "If I give up, now, then I know that I won't recover any more. And, I'll never have this opportunity, again, so maybe I should really try a bit harder/longer while I can..."

of course for people like curious george where time=money, its a double whammy of monetary and time inconvenience as well. as for me, its also partly my fault, as i did not expect this drive to fail "because its a wd" and a high end consumer drive line (caviar black) at that.

"Only back up the data that you care about..."

(The implication being that if you don't care about the data, why are you saving it?)

[ChaosLegionnaire]

Re: The hard drive failure thread

so a bad hard drive is just a monetary inconvenience?

i think its more than that. its a waste of time and energy as well. i have to spend time setting up the "drive recovery and rescue environment" and i have to wait for it to pull all the data off the drive before i can use my system again.

its also a waste of electrical energy as i have to leave my system on overnight so it can try to rescue and pull and reread all the data off the bad sectors. in the end, its how much time and energy u can expend to recover as much data as possible versus the loss of use of functionality of the machine u want to use while it performs the recovery.

of course for people like curious george where time=money, its a double whammy of monetary and time inconvenience as well. as for me, its also partly my fault, as i did not expect this drive to fail "because its a wd" and a high end consumer drive line (caviar black) at that.

[Curious.George]

Re: The hard drive failure thread

I do wonder now, should HDDs be considered disposable and have to keep it all backed up and/or RAIDded, so a bad hard drive is just a monetary inconvenience?

If you value the data on the drives, then you should plan on the drive failing at any time -- is your data worth more 3 years from now (when the drive is 3 years older) than it is, now? I.e., why would you consider backing up data on an older drive but not a newer one?

I keep at least two copies of everything on at least two different spindles. The really important things are backed up on different types of media.

Also, I regularly "verify/refresh" every instance of every file just so I know it's "still good (and accessible)" -- even optical media! (I have an automated system that alerts me to which volumes need to be mounted so that they can be checked "recently")

And, no RAID as that often limits your recovery abilities (what if the controller sh*ts the bed?) to whatever the controller decides is appropriate.

[eccerr0r]

Re: The hard drive failure thread

[Knock on wood]

I have a few 2TB drives, they all are still functioning.
The oldest is a WD green, and I have a refurbished WD (??color). The WD (??color) is reporting errors/bad sectors but I wonder if it's due to PSU problems as that machine keeps chunking PSUs. It is still functional, around 4+ years POH, and RAID1 with a Seagate 2T disk.

My WD Green 2T disk I couldn't stop Linux from load cycling until it was too late so it has a huge load cycle count. But it's over 5.5 years POH now. Fortunately no bad sectors.

The Seagate 2T disk (acquired new) seems okay but it does not have many POH as it's typically turned off along with the WD (??color) as they are RAID1.

I have a few other Toshiba 2T's as a RAID5 and they're average about 2 years old POH. Still doing okay.

I do wonder now, should HDDs be considered disposable and have to keep it all backed up and/or RAIDded, so a bad hard drive is just a monetary inconvenience?

[RJARRRPCGP]

Re: The hard drive failure thread

i also noticed that this live session of mini-xp started to stutter after the overnight write scan. strange... also, i have something like 200-300 slow sectors that take >600ms to be readable/writeable. i think im going to try to make the drive reallocate the slow sectors with >500ms response time using victoria.

I bet, if it was one of those Maxtor "slimeline" HDDs, you would keep seeing shit like that even when you told the utility to have slow sectors like that excluded!

And then likely a click-of-death later on...

[R_J]

Re: The hard drive failure thread

This may not be the problem but I have seen a lot of drives where the contacts between the board and the drive become tarnished and may be causing issues, cleaning them with an erasser my help.

[RJARRRPCGP]

Re: The hard drive failure thread

take a look at the reallocated sector phail from the smart data...

Looks like a variant of classic HDD failure and not because of a bad board/cache.

(On another note, I do know that certain Western Digital Caviar HDDs from pre-1998 or pre-1997, seem prone to bad sectors or a bad head and possibly general mechanical problems!)

(IIRC, I had one of the ones that I mentioned, a 1 GB, IIRC, which looked like it was manufactured in 1996 and they're would be a click of death beyond 238 MiB! The spinning was also unusually loud, IIRC, sounded like some friction against metal!)

[ChaosLegionnaire]

Re: The hard drive failure thread

well my wd caviar black 2tb kicked the bucket today. was trying to test a video card i bought on ebay so i booted up the system and while installing the video driver, the system was extremely sluggish copying and installing the files. i thought the sata cable was focking with me again and producing lots of ultra dma crc errors so i fired up crystal disk info to check. the cable was fine. instead i saw "caution" and it had one pending sector in smart, so it looks like the drive is failing.

im on ubuntu now running gnu ddrescue to pull all the data off it. i attached a screenshot of the smart data in linux. note that it has less than a year of poh on it. modern drives these days just dont last! sheesh!

[momaka]

Re: The hard drive failure thread

argg! i think that heavily vibrating hitachi p7k500 500gb is a bust. tried doing a write scan again and this time 250+ bad sectors popped up! however, the bad/slow sectors appear to be towards the end of the drive starting at around the 375gb mark. the beginning of the drive appears to be more or less fine. guess that drive will be relegated for testing purposes and os install testing only...

Just divide it into several partitions and use only the good ones. That's how I'm still using my 500 GB Seagate ST3500620AS I mentioned above. If the read and write tests are making it reallocate more sectors, but the drive is otherwise functional when you store files on it with your PC, then don't do anymore tests or benchmarks on it. My Seagate above is in the same boat.

that aside, i also want to talk about a hitachi micro hard drive failure in a 1.5gb creative muvo² portable mp3 player that occured several years ago. i bought that mp3 player in 2002 or 2003 during my national service days to kill time during slow days when i didnt have any ambulance calls.
...
what i didnt realise when i bought it in 2002/03 tho, was that those hitachi micro hard drives at the time were going for 600-700 bucks alone and the mp3 player that came with the micro hard drive was being sold for only 250-300 bucks.

That's crazy! I never heard of this.
I still have my Rio Carbon Pearl 6 MP3 player, which has a 6 GB Hitachi micro HDD in it. Actually, when I first bought the MP3 player back in 2006, someone in Best Buy must have dropped it, because it came with a dead HDD straight out of the box. I had to send it in for warranty to get it fixed. Ever since then, it's been running great - granted I put very very little use on it. I still like its design though, despite having only a monochrome screen.

[ChaosLegionnaire]

Re: The hard drive failure thread

take a look at the reallocated sector phail from the smart data...

[ChaosLegionnaire]

Re: The hard drive failure thread

argg! i think that heavily vibrating hitachi p7k500 500gb is a bust. tried doing a write scan again and this time 250+ bad sectors popped up! however, the bad/slow sectors appear to be towards the end of the drive starting at around the 375gb mark. the beginning of the drive appears to be more or less fine. guess that drive will be relegated for testing purposes and os install testing only...

that aside, i also want to talk about a hitachi micro hard drive failure in a 1.5gb creative muvo² portable mp3 player that occured several years ago. i bought that mp3 player in 2002 or 2003 during my national service days to kill time during slow days when i didnt have any ambulance calls.

the mp3 player served me well waay past my service until 2010, when it started to accumulate one too many drops to the ground by that time. i put the player in my shirt pocket and everytime i bend down to adjust/put on my socks or shoes or to pick stuff up off the ground, the player would sometimes fall out of my shirt pocket and onto the ground. first sign that the micro hard drive was failing was some slight clicking noises and skipping during playback. it suffered a few more drops until one day, it just refused to finish booting up on power on. player hung at the creative boot screen with a constant clicking coming from the micro hard drive. guess this should be in the dumbest electronic mistake thread as well...

i had my music collection on my pc so no loss there. what i didnt realise when i bought it in 2002/03 tho, was that those hitachi micro hard drives at the time were going for 600-700 bucks alone and the mp3 player that came with the micro hard drive was being sold for only 250-300 bucks. ppl were buying those creative mp3 players and tearing them apart to extract the micro hard drive on the cheap and used them in their other mobile devices like palm pilots etc. didnt realise all this until reading and hearing about it in a comp magazine and hearsay from colleagues who saw my mp3 player sometime in 2004.

makes me feel sad to see good ol' stuff like those costly hitachi micro hard drives kick the bucket from my carelessness at dropping my mp3 player. i was a noob back then and just tossed the mp3 player when it failed. now that i have turned a collector, in hindsight, i should have extracted the failed micro hard drive and just use it for display for posterity. my mp3 player now is a 16gb creative zen x-fi2 which uses flash mem for storage. no more screwing up the player from dropping it anymore. its endured several drops to the ground with no problems aside from more scuff marks... lol~

anyway, im starting to see the benefit of going with flash mem for mobile applications. hard drives are just too fragile to handle the rigors of mobile usage. if hard drives are going to become obsolescent and replaced with something better, its definitely going to start with the mobile arena. hard drives will still persist for fixed position, bulk storage usage tho. ok i will shut up now. this is a forum not a blog...

[Agent24]

Re: The hard drive failure thread

Spinning rust does have two operation modes: failing and failed.

The same goes for any storage device really, and in fact, all forms of computer hardware. Arguably it also applies to the entire known universe...

SSDs are not significantly better than anything else in the long run.

[Compgeke]

Re: The hard drive failure thread

Spinning rust does have two operation modes: failing and failed.

SSDs are getting to the point they'll probably outlive the average platter drive however they're still quite expensive for the same capacity. Keeping backups on more than one format in more than one place is the best for anything important.