Hard Drives Failing & CrystalDiskInfo

[freedomit]

Recently we have been using CrystalDiskInfo to quickly check and diagnose hard drive issues remotely. A couple of customer desktops we recently replaced the hard drives as CrystalDisk was reporting errors however i just scanned both with SeaTools and they both passed. Did we do the right thing replacing them?

Screenshots from both hard drives...

 

[wavey]

Any drive with reallocated sectors is a duff drive, so yes, replace.

 

[freedomit]

Any drive with reallocated sectors is a duff drive, so yes, replace.

That's what I thought but why does Seatools (which I trust more than CrystalDisk) report them as fine?

 

[nlinecomputers]

That's what I thought but why does Seatools (which I trust more than CrystalDisk) report them as fine?

Because your trust is misplaced. I trust CD more than Seatools. OEM tools are skewed towards not having an RMA.

 

[lcoughey]

It was a good move to replace. RMA and let Seagate deal with them. Of course, they will just send you back refurbished drives.

 

[Porthos]

Wasn't those DM drives riddled with failure issues?

 

[ohio_grad_06]

I personally am a fan of gsmartcontrol. Usually if that reports a drive as bad on the short test or has trouble completing the short test, out it goes especially if it's a customer system.

 

[Larry Sabo]

The problem is, some drive manufacturers try to intimidate you to discourage RMA of their drives by indicating that if the drive is found to not be faulty, they will return is and charge return shipping. At least that's what ADATA did regarding a faulty SSD of mine, which was "repaired" once then eventually failed completely and was replaced. A customer who has the identical drive (purchased and installed by me) is having speed issues and gSmartControl indicates ridiculously high Read Error/ECC Error rates. The ADATA Toolbox utility shows no problem. Some of these parameters are used differently by the manufacturer, so generic SMART tools don't yield reliable results. I like Wikipedia's helpful SMART parameter descriptions.

However, Reallocated sector/Reallocation event counts are rarely ambiguous, so you did the right thing.

 

[cypress]

I run book Crystal Disk Info and gSmartcontrol. If CD says pulls up caution and gSmatcontrol fails short test I immediately failed. Had one last week that CD failed but gSmartcontrol said it was fine. Recommended the client just replace drive anyway and upgrade to SSD. Drive was already 3 years old.

 

[FremontPC]

If the drive successfully reallocates, it likely will pass a SMART long or short test. If you have sectors pending reallocation, you might fail those tests. Indeed, the manufacturer's test is typically self-serving. They only care if the drive will perform for the warranty period. If the drive has reallocated sectors, in their eyes, that's what the pool of spare sectors is for and that's normal drive operation. That's my take on it anyway.

 

[ohio_grad_06]

Maybe we should find a reliable way of pushing drives over the edge. Not that I want to junk good drives, but if tools like gsmart or crystaldisk show them failing, and if the other tools are biased, then maybe there's a semi reliable way to finish killing the drive so then when you run their tool it will show it failed.

Again, I'm not wanting to like steal from manufacturers or anything, but how many times have we seen manufacturer tests show a drive good, another says no. We replace the drive and suddenly the machine is happy again.

 

[Mike McCall]

Recently had a client with a system that's 8-months old who's Seagate HDD was throwing astronomical read/write errors, but passed gsmartcontrol. Replaced the drive with a WD Black. Most Seagate drives I see throw high read/write errors no matter what I use to look at them. Gsmartcontrol even says that's not unusual for Seagate drives, though I don't know why that should be. The result (for me) is that I recommend replacing the drives if there's anything out of spec.

 

[freedomit]

Thanks everyone, just wanted to make sure i wasn't misdiagnosing and ripping of clients

 

[Larry Sabo]

Maybe we should find a reliable way of pushing drives over the edge.

I hope you're not serious?

 

[ohio_grad_06]

I was saying ONLY if the drive were failing other tests. Like for example, If a drive fails gsmart and crystal disk. But may not be a good thing to do. If I see a drive with an issue, I advise the person that we need to exchange it, and usually just buy a brand new drive. I don't really worry about the manufacturer's warranty anyway, I pay the extra 5 bucks or so for the 2 year replacement plan when I'm at microcenter getting a drive, that way if it goes south, I don't even have to talk to the manufacturer. I take it back and exchange it for a new unit, and if they want to go around with the manufacturer, that's their business. I don't have time to monkey around with doing that anyway.

 

[Larry Sabo]

I hear you about wasting time going the RMA route, but deliberately "pushing the drive over the edge" runs the risk of the customer accusing you of damaging a drive they considered still usable (albeit not particularly healthy). If word gets around, everyone with a drive that truly needs to be replaced is going to question your integrity, thinking maybe you did it to jack revenue. Anyway, that's the risk, as I see it.

 

[DataMedics]

Maybe we should find a reliable way of pushing drives over the edge. Not that I want to junk good drives, but if tools like gsmart or crystaldisk show them failing, and if the other tools are biased, then maybe there's a semi reliable way to finish killing the drive so then when you run their tool it will show it failed.

Dropping them while powered on, or even removing the PCB while it's powered on will usually do the trick.

 

[sapphirescales]

Some tools (and some technician's) don't mark a drive as bad if it has successfully reallocated sectors. And technically they're right. But a drive with reallocated sectors means the drive isn't healthy and is an indicator of possible imminent failure. Let me put it to you this way. I would never personally trust a drive with reallocated sectors with MY data. Why should you expect your clients to?

As for CrysalDiskInfo, that program is generally reliable. Basically any program that can read SMART data and give the proper values will work just fine. It's been a VERY long time since I've used Seatools, but back in the day they didn't actually report SMART data. They just had a short and an extended test you could run. Basically if the drive didn't die right then and there, they'd mark it as "Good." That's not a reliable test if you ask me.

 

[ohio_grad_06]

I hear you about wasting time going the RMA route, but deliberately "pushing the drive over the edge" runs the risk of the customer accusing you of damaging a drive they considered still usable (albeit not particularly healthy). If word gets around, everyone with a drive that truly needs to be replaced is going to question your integrity, thinking maybe you did it to jack revenue. Anyway, that's the risk, as I see it.

I can see what you are saying. Typically someone brings me a machine, I typically run gsmart short test on it, if it fails that or does not complete, or shows too many flags, I just recommend they replace the drive and quote what it would be for me to replace it. I don't do the RMA thing and all that anyway as I said, so there's not an incentive for me to do that. But I'm honest with people. If the machine is running slow and I see a potential problem with the drive, I'll tell them, look I can do a reinstall for X amount, but I recommend replacing the drive. I tell them it may potentially run for years, or it may run for a week. Then I give them the price if I replace the drive and leave it up to them to decide it.

On a customer unit, no I'm not going to deliberately destroy a drive just to make a profit. Now if it was a drive in my personal system and I'm dealing with a manufacturer for something for myself, might be a different story, but on customer systems, a no no. I was thinking out loud to. Maybe too much.

 

[Diggs]

Any drive with reallocated sectors is a duff drive, so yes, replace.

For the sake of discussion I would tend to differ with this. I've seen quite a few drives (laptops) that were plopped down on a table while writing and produced sector errors. After re-allocating (and re-educating the owners) these drives/laptops went on to serve long (years) lives.