Hard disk repairing/recycling

[x5ga]

Ugh, it seems that (another) one of my hard disks has a bit of trouble accessing the data that's written on it. It's a single-partition 320GB SATAII Western Digital Caviar Blue (WDC3200AAKS) with 140GB of mp3/flac/ape files (all my perfectly sorted music collection), so it would be somewhat BAD if all the data would get corrupted.

The drive gets detected perfectly in the BIOS, but once you start reading the data (ex: playing one of the songs or even navigating through the folder structure), it stalls for a long time. Then it usually reads the data correctly, but the 1-2minutes lag is more than annoying. The other problem the drive has is that Windows would stop detecting it after it's been used for a while (it just disappears from the drives list). It's like *magic* and the drive isn't listed anymore. On reset, it's back again, only to disappear once more after a few hours.

The cables are fine, checked and replaced them twice. The controller is good also. AHCI is disabled, and the drive isn't part of a RAID array.
Ran chkdsk on the drive, and it marked 72K as "bad", no other problems. SMART says the drive status is OK (lol, SMART is actually DUMB most of the time anyways). No weird noises coming out of the drive either.
My PC specs (the one with this hdd in it): Gigabyte GA-965P-DS3, Core2Duo E6320@3GHz (OC'd), 5GB RAM, Ati Radeon HD4670, OS: Windows 7 x64.

So... what I want to know is the following:
- Is the drive gonna completely die on me soon? (yeah, it probably will, won't it... )
- Can some software (ex: SpinRite) help with the problem(s), at least temporarily? (1-2 months 'till I get a new drive)
- Will a complete reformat / some software / voodoo magic save the drive? (again, just for 1-2 months)

Also, will changing the electronic part of the drive (the PCB&stuff) help? I have an almost identical hard drive laying around (it's another WDC3200AAKS, only ~1 year newer than the near-death one, but has the same technical specifications), but I'm not gonna cannibalize it without a good reason, since it's better to lose 1 drive than 2

Thank you

[kureshii]

SMART Status = ok says pretty much nothing.

Check reallocated sector count and temperatures in SMART, and other key indicators (do a quick google), that tells you much more useful info than "SMART Status: ok". SMART isn't dumb (it's not a "your-drive-will-die-on-15-Nov-2010" tool), unless programs attempt to summarise its myriad data in dumb ways.

What to do now? Use the drive as little as possible of course. And try to find a replacement soon. Make backups, quick.

[x5ga]

SMART is reporting some weird stuff with the drive... and I'm using both O&O driveLED and CrystalDiskInfo, both report the same thing. The only parameters that are not "ok" are 'unrecoverable error count', which is 27(instead of 0) and 'unstable sector count', which is 28(instead of 0). The weird thing is that the drive apparently has 870h total power-on time and 1547 start-ups. That's ... wrong. The drive is ≈3 years old and has been powered on for at least 2 years. Heck, the system has a 3rd drive which I added ≈1 year ago and that drive has 3800h uptime. So, SMART can apparently be dumb sometimes. Still, the status is still 'green', even though those 2 parameters are not normal.

So I guess I better make backups and throw the drive away?

[kureshii]

1) Most programs report the SMART Status based on threshold values, which are vendor-specified. So you should look at the raw values for key indicators, since sadly manufacturers tend to inflate these threshold values and there doesn't seem to be anything preventing this. But the raw values should at least be accurate. Even if they don't, various other factors can yield broken threshold values (bad drivers, unspecified threshold values etc).

2) AFAIK Windows' default power setting is to spin down idle drives. Since you mentioned a single-partition drive and made no mention of your system or other data, I assume that drive is being used mainly for music and other non-system stuff. In that case, 870h might not be so unreasonable. Although it is possible that other reasons account for this, e.g. drive RMA, incorrect calculation or simply wrong SMART data (yes, this happens sometimes, often because of faulty firmware).

tl;dr Stop caring so much about "overall SMART status" and learn to read the key indicators instead. A quick googling never hurts. This is a good introduction to SMART, whether you use HD Sentinel or not.

The error counts certainly don't look good. Do check the temperature data and see what range of temperatures the drive has been running at (although it doesn't say anything by itself, drives generally tend to fail earlier if left to run at really high temperatures for long periods of time).

You can keep using the drive, but you should back up critical data regularly (twice daily if you want to be safer), and mainly use the disk for noncritical stuff. Basically, use it as an occasional external, for passing your friends stuff or just to hold some data temporarily. Use the drive like you expect it to die any time. If you're lucky either it holds out beyond your expectations, or it dies really early so you get to RMA it.

[x5ga]

hehe, thanks, that article was a good read... though the raw values are ok, except those two which I mentioned earlier. By default, Windows does spin down the drive, but I have music playing in the background off that HDD for a long time (almost 24/7), so, at least the uptime (yes, the raw data) is wrong, in my opinion - not that this makes the drive less likely to fail. I'll just backup the data somewhere and then copy some non-critical data onto the disk and put it in a closet somewhere safe. Maybe I'll try formatting it and running spinrite and/or HDDregenerator, it probably can't hurt...

Thanks for your feedback

[Pentium100]

Windows does spin down the drive, but I have music playing in the background off that HDD for a long time (almost 24/7), so, at least the uptime (yes, the raw data) is wrong, in my opinion - not that this makes the drive less likely to fail.

AFAIK, Windows has predictive cache, so in your case it would read the music file that's playing and spin down the drive if the file is long enough.

I usually set Windows to never spin down the drives. One of my older hard drives has 44496h of usage and only 332 power cycles. A newer drive has 36626 hours and 123 cycles.

I'll just backup the data somewhere and then copy some non-critical data onto the disk and put it in a closet somewhere safe. Maybe I'll try formatting it and running spinrite and/or HDDregenerator, it probably can't hurt...

Good idea! Run MHDD then Spinrite. MHDD can scan (and erase) the drive faster than Spinrite, which tries to restore the data that's on the drive. MHDD in scan mode should report to you how many bad sectors this drive has. A few of them is not a problem, but watch carefully to see if the number grows with use.

[x5ga]

Windows does spin down the drive, but I have music playing in the background off that HDD for a long time (almost 24/7), so, at least the uptime (yes, the raw data) is wrong, in my opinion - not that this makes the drive less likely to fail.

AFAIK, Windows has predictive cache, so in your case it would read the music file that's playing and spin down the drive if the file is long enough.

I usually set Windows to never spin down the drives. One of my older hard drives has 44496h of usage and only 332 power cycles. A newer drive has 36626 hours and 123 cycles.

Even with the predictive cache, it wouldn't account for the 10000+ hours the drive should have had, and I disable the spin-down too... The drive must be reporting its uptime erroneously...

I'll backup the data first thing in the morning and then run MHDD and spinrite, hopefully it'll magically make the drive OK for a while at least I'll keep track of the bad sectors also. This is the first Western Digital drive that has caused me problems, and I have ≈8 of them running, one of them has 110000h uptime and has zero problems. Maybe this one was from a 'less-than-good' batch or something.

[Pentium100]

one of them has 110000h uptime and has zero problems. Maybe this one was from a 'less-than-good' batch or something.

Nice, 12.5 years. That long ago, I could not leave my PC on all the time...

Also, the drive may report something else and not hours in that SMART attribute. Maybe it's 870 days Check it today and tomorrow and see how much it changes...

1547 power cycles is still a lot, on average, your drive was turned on at once every 17 hours during those 3 years. Normal for regular PC, but not for one that's on 24/7.

[fohfoh]

I have a similar drive that does the same thing. However, mine was in an external enclosure and writing data when it was knocked off the table by my dog.

It "works" now, but not really. I merely use it in my HTPC to drop files to play off of because it's faster than the 1TB WD Green installed. I don't trust it, but it was my fault that it was like that... Not WD's fault.

I concur though. SMART means nothing. I've done disc scans etc that say that WD is perfectly fine, but it's not. Even trying to install ubuntu on it yields poor results. (Failure)