The hard drive failure thread

[RJARRRPCGP]

Re: The hard drive failure thread

Not at all. Looks pretty clean, actually.

Only "unclean" value I see is Ultra DMA CRC Error Count

I thought I saw the same issue with Seagate Momentus HDDs...

[RJARRRPCGP]

Re: The hard drive failure thread

Western Digital Caviar 1GB drive, WDAC21000-00H. Had a Mfg. date of 1996-02-29. Finally failed in 2004 due to Click of Death. I called it the Leap Year drive.

I had one of a similar model, if not the same model and in 2003 or the forth quarter of 2002, fails with a click of death, if you try to go beyond 238 MB.

[momaka]

Re: The hard drive failure thread

Does pic look like a drive with problems?

Not at all. Looks pretty clean, actually.

Only "unclean" value I see is Ultra DMA CRC Error Count, which typically results from a bad SATA/IDE cable or bad PCB contacts with the headstack / head amps. But since bad SATA cable is unlikely to be the issue and since you also already fixed (tinned) the PCB contacts, most likely any value left in the UDMA CRC Error Count is from before, when the drive had the dirty/tarnished PCB contacts. IIRC, UDMA CRC Error Count can never be reset, even after zero filling. So probably nothing to worry about.... but do watch it. If the UDMA CRC Error Count keeps going up, the drive likely still has problems with the PCB contacts or possibly the headstack itself.

[pfrcom]

Re: The hard drive failure thread

Could be that the HDD was just intermittent before too??

I had put it aside with a note saying undependable

Anyway, while walking dog, I thought about differences between my testing setup and the dc7800 - namely IDE vs AHCI

So I changed the dc7800 to IDE and did another full scan, which came up clean

Then I continued to set up triple boot - XP, Win7 & MXlinux - which all went well, although I was nearby, not letting it go to sleep

I did Google "WD2500JD sleep problems" and "WD2500JD ahci problems" in case of a firmware update, without finding anything

Does pic look like a drive with problems?

[momaka]

Re: The hard drive failure thread

Well, at least it "lived" for a bit longer. Perhaps the full scan + full erase was too much for it? IIRC, those particular drives tend to run hot - especially in an enclosure. Not sure if that plays a part or not.

Also, it's interesting to note that, based on your description, the drive is still trying to power on / calibrate but fails somewhere halfway there? On that note, I wonder if putting it in different orientation will do anything. Just makes me wonder if the brief amount of time you got it working was a coincidence or really a result of cleaning the pads. Could be that the HDD was just intermittent before too??

[pfrcom]

Re: The hard drive failure thread

clean the head amp PCB pads, and often even tin them with Leaded solder

tried similar with a WD2500JD

Used Deoxit Red (on overnight) then Deoxit Gold

To remove the Deoxit, I cleaned the pads with a typist's eraser pencil (from back in the day), then tinned them with 60/40, finally removing excess solder with Chemtools wick

Wasn't that optimistic, but pleasantly surprised to hear Bzzzt heads loading noise instead of clack-clack... with drive being detected by BIOS

Then went through WD Data Lifeguard Full Scan, Full Erase and Full Scan again without missing a beat

Finally put it in a HP dc7800 to use instead of a bigger drive, and left it installing Win7

Came back to the screen for entering user & machine names and after entering something, CLACK---zhzhzh---CLACK---zhzhzh... instead of cleanly coming out of sleep mode

Damn thing - <7200 hours on it

[momaka]

Re: The hard drive failure thread

I recently tried similar with a WD2500JD

Went from:
CLACK----CLACK----CLACK----CLACK

To:
clack-clack-clack-clack-clack-clack

Oh, I have one of those as well. This was one of the WD HDDs on which cleaning the contacts did *not* work (HDD still clicked the same afterwards.)

So tinning the pads will by no means revive every bad WD HDD back. But it's one less thing to fail (and still worth a shot if nothing else works), so that's why I do it.

[pfrcom]

Re: The hard drive failure thread

clean the head amp PCB pads, and often even tin them with Leaded solder

I recently tried similar with a WD2500JD

Went from:
CLACK----CLACK----CLACK----CLACK

To:
clack-clack-clack-clack-clack-clack

Used Deoxit Red (on overnight) then Deoxit Gold

Didn't think of tinning though - hmmm, has joined list of stuff to try

[momaka]

Re: The hard drive failure thread

not sure what happened to the drive on that day to cause it to crap out... maybe the psu was unstable on that day as i noticed on the pc health status screen in the mobo bios setup screen that the 12v rail was only 11.4v. could be unstable power causing the drive to crap out esp. when writing data and malfunction. when power was stable again, the data written to the drive was stable and fine thus it seems to work again. could be a dirty sata power connector or bad psu modular cables making intermittent contact.

No, that's not how PSUs and power work at all!

For starters... NEVER trust the motherboard voltage sensors - at least for the ones that directly measure the PSU's rails. The only voltage sensor I somewhat trust are the CPU V_core... and even then, I always double-check it with a multimeter. On that note, it's worthwhile to mention that crappy/cheap meters can also show garbage readings. Using two of my cheapo HFT meters to check the 5VSB on an Antec PSU I was fixing... one read 4.75V and the other 5.15V. So is the voltage on the 5VSB high or low? I then took my "more serious" Radio Shack DMM and checked with it - the 5VSB was 4.99 to 5V right on the spot. I proceeded to confirm the HFT meters were showing garbage by using a known 2.5V and 5V reference voltage source.

Also bad power isn't going to necessarily result in high Raw Read Error Rate. If anything, I think that's just due to the fact that you have a WD Green drive, which are known for their problems / bad reliability issues.

More to the point: I'd lift the PCB from the drive and check the head-amp / actuator arm pads on the board. On WD drives, they are known to tarnish / oxidize after some time due to bad/cheap tinning of the PCB. This goes as far back as the first WD SATA HDDs and even affects some of their IDE HDDs close to the beginning of RoHS era. So when I get an old WD drive, I always pull the PCB now, clean the head amp PCB pads, and often even tin them with Leaded solder. They'll never tarnish again afterwards. Also managed to revive two old WD SATA HDDs that way when they weren't showing up in BIOS at all and clicking/seeking like mad.

i also hadnt used that system for several months, so it could be that the jp-ce turd caps in the enermax liberty 500w degraded while sitting unenergized. when power was applied, the caps were able to reform but not without providing degraded filtering performance for a short while which caused unstable power. my enermax liberty has white silicone glue not the brown conductive one so its not stuff getting shorted inside the psu making the power unstable.

Good theory, but no. A few months in storage is not really a big deal for caps. If the caps failed in that time... it means they are either crap and were probably going to fail anyways. Either way, they can't get "better" afterwards. So the whole thing "HDD was unstable before but is stable now" likely has nothing to do with the PSU power.

On that note... the tan/brown "conductive" glue doesn't cause intermittent problems either. It is only dangerous on the primary side (and sometimes very rarely on the secondary side), where after some time it will become conductive enough to mess up driving signals in the PSU... or worse, bias a MOSFET Gate when it shouldn't be and cause a major kaboom. Either way, both of these cases will directly result in a non-working PSU, not a "was bad before, is OK now" scenario.

i guess this is another cautionary tale not to use junk psus with modern hard drives. guess i should stop using that enermax liberty until i get rid of those jp-ce turd caps...

I wouldn't call that Enermax PSU junk. Granted the PCE caps in there are pretty old and probably are due for a change now... but I still find it highly unlikely this was the cause of your HDD malfunctioning.

[Curious.George]

Re: The hard drive failure thread

been reading and verifying the sectors on the drive and no errors. tried re-writing the lba numbers again with different date and time stamps again and still no issues. not sure what happened to the drive on that day to cause it to crap out... maybe the psu was unstable on that day as i noticed on the pc health status screen in the mobo bios setup screen that the 12v rail was only 11.4v. could be unstable power causing the drive to crap out esp. when writing data and malfunction. when power was stable again, the data written to the drive was stable and fine thus it seems to work again. could be a dirty sata power connector or bad psu modular cables making intermittent contact.

11.4V is -5%. Note that the DISK (and controller) really only care about what the supplies are LOCALLY, not "someplace else".

# dd if=/dev/rawdevicesource of=/dev/rawdevicedestination bs=1024k
# cmp /dev/rawdevicesource /dev/rawdevicedestination

neither should throw ANY errors. And, should hammer on the drive pretty hard (though all sequential accesses).

[ChaosLegionnaire]

Re: The hard drive failure thread

well... over the weekend, i performed some read and write tests on the caviar black 2tb. i first used the write lba numbers test in victoria which writes date and time stamp data on the drive's sectors as well and the write operation over the entire drive went fine and the drive did not reallocate any sectors.

the drive seems to be still in perfect health aside from the raw read error rate, most likely racked up from trying to constantly re-read bad sector data during the drive recovery operation with gnu ddrescue in linux.

been reading and verifying the sectors on the drive and no errors. tried re-writing the lba numbers again with different date and time stamps again and still no issues. not sure what happened to the drive on that day to cause it to crap out... maybe the psu was unstable on that day as i noticed on the pc health status screen in the mobo bios setup screen that the 12v rail was only 11.4v. could be unstable power causing the drive to crap out esp. when writing data and malfunction. when power was stable again, the data written to the drive was stable and fine thus it seems to work again. could be a dirty sata power connector or bad psu modular cables making intermittent contact.

i also hadnt used that system for several months, so it could be that the jp-ce turd caps in the enermax liberty 500w degraded while sitting unenergized. when power was applied, the caps were able to reform but not without providing degraded filtering performance for a short while which caused unstable power. my enermax liberty has white silicone glue not the brown conductive one so its not stuff getting shorted inside the psu making the power unstable.

since it seems to be working fine now, i guess i'll put it back into service by cloning all the data off the hitachi 2tb drive i rescued the data to back to the cav black 2tb drive. i guess this is another cautionary tale not to use junk psus with modern hard drives. guess i should stop using that enermax liberty until i get rid of those jp-ce turd caps...

[RJARRRPCGP]

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 !

That one makes me suspect a shit HDD board or a shit component on the HDD board.

[RJARRRPCGP]

Re: The hard drive failure thread

When I used to work in a electronics repair shop, laptops and PS3s with failed Toshita HDDs were common sight.

In the current period, I have a feeling that I will see far more failed Seagate Momentus HDDs than Toshiba HDDs!

I likely had to condemn two Seagate Momentus HDDs already! (even when not sure if I did or not)

[RJARRRPCGP]

Re: The hard drive failure thread

idle in the mid forties, under load they'll burn your hand!!!

Like an Intel CPU, at least with the stock heatsink, LOL

[RJARRRPCGP]

Re: The hard drive failure thread

Sometimes I wonder, why we don't go back to software based bad sector mapping... it was the norm years ago.

Yes, it was 22 years ago. (Or at least 21 years ago.)

[eccerr0r]

Re: The hard drive failure thread

Sometimes I wonder, why we don't go back to software based bad sector mapping... it was the norm years ago.
Then again, what was bad is that the bad blocks were written on the drive itself and you had to copy that data in the OS, and undoubtedly the number of errors on a modern hard drive is many times fold of those old 10MB hard drives.

I recall having a few hard drives where the error map said

none

and be glad I don't have to type anything in the OS.

This does not help against head crashes and other drive failures of course, just media issues.

[RJARRRPCGP]

Re: The hard drive failure thread

Western Digital
WD1600BEKT
18 Aug 2009 (under warranty until 2014, hooray)
1420 POH (yes, 1,420)
SMART reads "Pending sectors", system gets stuck in a reboot loop. Error scan shows this-

Seems to act much like one of those Maxtor "slimeline" HDD models! Those Maxtors, (looking at you 6E030L0!) have a known issue where the firmware epically fails to remap sectors.

[Curious.George]

Re: The hard drive failure thread

[predictive powers]

Oh, but I do.

Then you're in an envious position! I have to keep past projects on the chance that a client will want some tweak at a later date. So, I freeze my entire workstation state at the time I deliver a project (so I have all the tools and their configuration preserved, as well). Otherwise, I'm at the mercy of the client to have retained a copy of my "deliverables" that I can work from. He'll often have little incentive to do so as he will have made modifications and want me to "inherit" those modifications in the changes that I make (translation: fix his f*ckups!)

My music and some of my pictures are really the only two things I care about - and hence the only two I backup the most.

I can probably recover most of my music collection from other sources, if need be; they aren't "unique to me". And, I only take photographs of things (not people) so there's no real attachment there. If I could never-again see a photo of a prototype that I built 40 years ago I wouldn't shed any tears...

Lately, I've also started to backup my data sheet library, as I've spend a considerable amount of time to organize it, and some data sheets weren't exactly easy to find. Same with certain drivers for hardware that I still own & use.

Datasheets get captured with the project to which they applied. So, I don't have to chase down a particular part number in a "library". Instead, the parts used on project X are captured along with the tools and work done on project X. This also makes it easier for me to remember where to find a particular datasheet: "Which project did I use that in?"

I keep an archive of "Kit" with folders for things like Computers, Appliances, Automobiles, etc. Within, for example, Computers, you'd find a folder for each manufacturer (of devices that I've owned or own currently). And, under those, folders for each device -- containing drivers and "build notes" archived from each time I built a system on that particular hardware.

E.g., today I'm adding stuff that I've found when researching a Nest thermostat I rescued a few weeks ago so everything "Nest Thermostat" related sits in a single folder.

As for any projects, receipts, tax documents and etc. - I couldn't care if these went with the devil one day. I back them up simply because they don't take that much space.

I keep all of my tax records for my business as the cost of NOT having them in an audit far outweighs the effort to retain them. It's also handy when I need to sort out how long ago I bought a particular item -- it will be recorded in my purchases for SOME past year!

Movies and game files do take up a lot of space, so these I don't really backup either, unless it's something that I use often and I know is hard/impossible to find now (came in point: some old game mods that have long been gone from the internet and no one cares about anymore.)

Any games that I had are now pretty old (Doom vintage) and don't see any real play. I keep ISOs of all of the CD's as well as the PSP disks. But, they could easily disappear without me (personally) regretting it. OTOH, I keep ROMs for the various video (arcade) games I've owned -- even if I've already given away the arcade piece (there's so little to be SAVED/reclaimed by erasing them that its not worth the effort!)

[momaka]

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.

Oh, but I do.

My music and some of my pictures are really the only two things I care about - and hence the only two I backup the most. Lately, I've also started to backup my data sheet library, as I've spend a considerable amount of time to organize it, and some data sheets weren't exactly easy to find. Same with certain drivers for hardware that I still own & use.

As for any projects, receipts, tax documents and etc. - I couldn't care if these went with the devil one day. I back them up simply because they don't take that much space. Movies and game files do take up a lot of space, so these I don't really backup either, unless it's something that I use often and I know is hard/impossible to find now (came in point: some old game mods that have long been gone from the internet and no one cares about anymore.)

[Curious.George]

Re: The hard drive failure thread

Roughly on par for Windows 7 Embedded.

The difference is that XPe is "just a client". You wouldn't let it provide core services for a network (BIND/NTPd/FTPd/NFSd/BOOTPd/TFTPd/DHCPd/etc.) -- or include development/debugging tools to be an extensible platform -- "out-of-the-box".

(I repurpose XPe appliances to run NetBSD diskless/headless... what value does XPe have without a keyboard and display?)

We already know the sources for XPe aren't even available, let alone "included".

Nor is XPe "free" and independently maintainable (once EoL'ed).

And, of course, you don't have to think about rebooting it as its sole "fix" when it f*cks itself!