On hard drives health and Disk Utility...

[iRock1]

Hi.

I was wondering, what's the best option out there? when it comes to check the hard disks health? (I mean internal and external, i.e. primary disk and attached storage.)

I've read that Disk Utility does the job pretty well, though to be honest I'm totally a noob when it comes to hard disks. Is the “verify disk” option going to really tell me if a disk is going to die soon? I want to be extra careful when it comes to handling information on those drives. Or should I use an additional software to make 100% sure that my disks are working ok? (Obviously I know it's impossible to know when a hard disk will fail, but we are talking about probabilities here.)

And one last thing. Why can I verify my primary disk from Disk Utility, but not my external ones? I mean, I can only verify each of the partitions on them, but no the actual disks.

 

[monokakata]

I read about this product:

http://scsc-online.com/Scannerz.html

on this thread:

https://forums.macrumors.com/threads/1741407/

and went ahead and bought it.

So far I've run it on two 3 TB disks. It takes a long time to complete its scan. One drive was flagged as potentially having problems. This is part of the report:

"The initial surface scan test passed. The initial linear surface scan RPI has an acceptable value. The initial surface scan ratio RPI has an abnormally high value. This implies that regions of the disk are taking considerably more time responding than others, which may imply the drive is in the process of developing problems."

So, I'll watch this drive and probably test it again in a week to see what's happening. If it's still laggy then I'll start using it only for non-critical stuff.

Two tests, two disks -- I haven't had much experience with it. But so far, I like this app.

 

[iRock1]

I read about this product:



http://scsc-online.com/Scannerz.html



on this thread:



https://forums.macrumors.com/threads/1741407/



and went ahead and bought it.



So far I've run it on two 3 TB disks. It takes a long time to complete its scan. One drive was flagged as potentially having problems. This is part of the report:



"The initial surface scan test passed. The initial linear surface scan RPI has an acceptable value. The initial surface scan ratio RPI has an abnormally high value. This implies that regions of the disk are taking considerably more time responding than others, which may imply the drive is in the process of developing problems."



So, I'll watch this drive and probably test it again in a week to see what's happening. If it's still laggy then I'll start using it only for non-critical stuff.



Two tests, two disks -- I haven't had much experience with it. But so far, I like this app.

Just curious, what do you get checking the same drive with Disk Utility?

 

[monokakata]

No messages -- DU thinks it's fine, because the file system and the other things that DU checks are in order.

And at the moment, it's working as it should -- from the point of view of a user.

Scannerz is saying "you may see a problem down the road." It's checking an an entirely different (lower) level.

That's why over the weekend I'll run Scannerz again. If it says again that it didn't like the response it got, then I'll move the disk to non-critical use.

 

[Weaselboy]

Hi.

I was wondering, what's the best option out there? when it comes to check the hard disks health? (I mean internal and external, i.e. primary disk and attached storage.)

I've read that Disk Utility does the job pretty well, though to be honest I'm totally a noob when it comes to hard disks. Is the “verify disk” option going to really tell me if a disk is going to die soon? I want to be extra careful when it comes to handling information on those drives. Or should I use an additional software to make 100% sure that my disks are working ok? (Obviously I know it's impossible to know when a hard disk will fail, but we are talking about probabilities here.)

And one last thing. Why can I verify my primary disk from Disk Utility, but not my external ones? I mean, I can only verify each of the partitions on them, but no the actual disks.

Take a look at SMART monitoring utilities like DriveDX. As a drive starts to have errors they will be reported in the SMART status with these utilities.

In my experience, DU's verify disk mostly shows problems after they have already occurred and need repair.

 

[TheBSDGuy]

SMART testing only detects errors after the fact, which is why it isn't very good. Scanning tools like Scannerz can find problems in areas in regions that haven't been accessed yet, and if I'm not mistaken it already has SMART evaluation built into it anyway.

The Google study on hard drives essentially said that SMART testing wasn't much of a predictor of failure.

 

[Weaselboy]

SMART testing only detects errors after the fact, which is why it isn't very good. Scanning tools like Scannerz can find problems in areas in regions that haven't been accessed yet, and if I'm not mistaken it already has SMART evaluation built into it anyway.

The Google study on hard drives essentially said that SMART testing wasn't much of a predictor of failure.

I have no experience with Scannerz, but I am guessing you did not read the DriveFX web page before commenting. DriveDX does the full media scanning you mentioned Scannerz does.

You really should reread the Google study. I agree SMART is imperfect, but the Google study showed 60% of the drives that failed were showing SMART errors.

 

[ZVH]

I wrote the original article on the list of drive testing software. When I found out about DriveDX, I put it on the list. I later removed it. Why?

If you right click on the DriveDX binary, and select "Show Package Contents", double click on "Contents" followed by "Resources" you will see a binary named "smartctl" sitting inside the package. "smartctl" is the control program for smartmontools, which is listed in my article.

Basically what they've done is taken a free software package, wrapped a GUI around it, and then called it their own. If the DriveDX guys had put that information out publicly, I would have no problem with it, but to hide it down in the bowels of an application with no acknowledgement is, IMHO, unethical, and quite possibly illegal, at least in the U.S. anyway. There was another product on the market that I was aware of that did the exact same thing, but I knew about it and never put it on my list of drive testing software.

Another thing I didn't like is the fact that a lot of SMART features aren't present or built into a manufacturers implementation of SMART, but the manual makes it sound like all SMART features are supported. SMART features are only supported if the manufacturer decides to supports them, and in many cases the items they support are minimal. Surface scanning, for example is very, very often not implemented. The manual implies it will do those tests, almost unconditionally, when in fact a lot of drives are totally incapable of it.

People interested in SMART testing should get the REAL smart program, smartmontools at:

http://www.smartmontools.org/

Regarding SMART reliablity, see:

http://storagemojo.com/2007/02/19/googles-disk-failure-experience/

From there they quote:

How smart is SMART?
Not very, as Google found, and many in the industry already knew. SMART (Self-Monitoring, Analysis, and Reporting Technology) captures drive error data to predict failure far enough in advance so you can back up. Yet SMART focuses on mechanical failures, while a good deal of a disk drive is electronic, so SMART misses many sudden drive failure modes, like power component failure. The Google team found that 36% of the failed drives did not exhibit a single SMART-monitored failure. They concluded that SMART data is almost useless for predicting the failure of a single drive.

In my opinion, SMART is just another check, if treated as just another check, but for the love of God don't think it's the be-all end-all in testing, far from it.

 

[Weaselboy]

I wrote the original article on the list of drive testing software. When I found out about DriveDX, I put it on the list. I later removed it. Why?

If you right click on the DriveDX binary, and select "Show Package Contents", double click on "Contents" followed by "Resources" you will see a binary named "smartctl" sitting inside the package. "smartctl" is the control program for smartmontools, which is listed in my article.

Basically what they've done is taken a free software package, wrapped a GUI around it, and then called it their own. If the DriveDX guys had put that information out publicly, I would have no problem with it, but to hide it down in the bowels of an application with no acknowledgement is, IMHO, unethical, and quite possibly illegal, at least in the U.S. anyway. There was another product on the market that I was aware of that did the exact same thing, but I knew about it and never put it on my list of drive testing software.

I'm not following your outrage here at all. smartctrl is open source software and there is absolutely nothing unethical or illegal about using it in commercial software. The use of smartctrl is disclosed in the Acknowledgements section of the help menu where this information is often found. It is not hidden "in the bowels of the application."

I get that it would make you feel better if they blasted that across their web page, but they are under no obligation ethically or otherwise to do so.

This is how open source works. There are many commercial applications that are built on or use open source projects and there is nothing wrong with that. The very popular cloning utility Carbon Copy Cloner includes the rsync binary package. Go look at the CCC homepage and tell me if you see that disclosed anywhere? You won't find it because it is not there. If you dig around you can find it disclosed here. Should we all stop using CCC because they did not advertise this well enough for you?

Here is the excerpt from the acknowledgements.html file clicking on that menu opens:

Acknowledgements

Portions of this software may utilize the following copyrighted material, the use of which is hereby acknowledged:

smartmontools, Copyright (c) 2002-14 Bruce Allen, Christian Franke, www.smartmontools.org.
smartctl executable binary is included with this software.

### AUTHORS ###
This code was originally developed as a Senior Thesis by Michael
Cornwell at the Concurrent Systems Laboratory (now part of the Storage
Systems Research Center), Jack Baskin School of Engineering, University
of California, Santa Cruz. http://ssrc.soe.ucsc.edu/

...

Maintainers / Developers:

For more details please see file 'smartctl_copyrights.txt' (in app bundle resources directory).

Regarding SMART reliablity, see:

http://storagemojo.com/2007/02/19/googles-disk-failure-experience/

From there they quote:

"They concluded that SMART data is almost useless for predicting the failure of a single drive."

That is nothing but hyperbole from the author of the Storagemojo article and it is certainly not a quote from the Google paper. Go read the actual Google study here and you will see that is not at all what it said. The study showed SMART is imperfect just like I said, but the bottom line is SMART was showing errors on 60% of the failed drives in the Google test, so obviously is it not "useless". The word useless never appears once in the entire Google paper.

In my opinion, SMART is just another check, if treated as just another check, but for the love of God don't think it's the be-all end-all in testing, far from it.

I don't think anybody here said anything like that.

 

[ZVH]

CCC openly and clearly acknowledges that they use not only rsync, but a specific version of it. CCC did it right.

You have much more control over smartctl by using the command line version, and your telling the OP to go out and fork over yet another 20-25 bucks for something that he can get for free, and something that's already built into Scannerz, which he already owns.

From the Google article:

Our analysis identifies several parameters from the drive’s self monitoring facility (SMART) that correlate highly with failures. Despite this high correlation, we conclude that mod- els based on SMART parameters alone are unlikely to be useful for predicting individual drive failures. Surprisingly, we found that temperature and activity levels were much less correlated with drive failures than previously reported.

Given the lack of occurrence of predictive SMART signals on a large fraction of failed drives, it is un- likely that an accurate predictive failure model can be built based on these signals alone.

We conclude that it is unlikely that SMART data alone can be effectively used to build models that predict fail- ures of individual drives.

Despite those strong correlations, we find that failure prediction models based on SMART parameters alone are likely to be severely limited in their prediction accuracy, given that a large fraction of our failed drives have shown no SMART error signals whatsoever

 

[Weaselboy]

CCC openly and clearly acknowledges that they use not only rsync, but a specific version of it. CCC did it right.

Oh please. CCC has the info buried on their web page and that is better than an acknowledgment right in the app? Just be honest and admit you did not know that acknowledgement was there in the app and you made mistake. It is obvious by your comment that the info is only hidden in the "bowels of the app" that you missed the clear acknowledgement in the help menu. You are raging about a complete non-issue here.

You have much more control over smartctl by using the command line version, and your telling the OP to go out and fork over yet another 20-25 bucks for something that he can get for free, and something that's already built into Scannerz, which he already owns.

Please show me where the OP iRock1 said he already owns Scannerz. If you have a better and free way to help out the OP, by all means post it up for them. I'm sure not trying to discourage that.

From the Google article:

Like I have said repeatedly, I agree SMART is imperfect and I actually have read the Google study. Nowhere in the Google study does it say SMART reporting is "useless" like the quote you posted.

 

[SaSaSushi]

I wrote the original article on the list of drive testing software. When I found out about DriveDX, I put it on the list. I later removed it.

And you removing it from this list of yours is relevant to what exactly? As if they needed your endorsement.

I'm not following your outrage here at all. smartctrl is open source software and there is absolutely nothing unethical or illegal about using it in commercial software. The use of smartctrl is disclosed in the Acknowledgements section of the help menu where this information is often found. It is not hidden "in the bowels of the application."

Exactly correct. Smartmontools is also used by SmartReporter (as well as every other OS X SMART app), which apparently is on his list. How random.

----------

Take a look at SMART monitoring utilities like DriveDX.

+1 for DriveDX. I've mentioned before in other threads on this topic that the app has already saved me data by accurately predicting the failure of a backup HDD I had no other indications of problems with.

It's not cheap as SMART utilities go, but it's well worth its price IMO.

----------

Like I have said repeatedly, I agree SMART is imperfect and I actually have read the Google study. Nowhere in the Google study does it say SMART reporting is "useless" like the quote you posted.

Just like it says in the StorageMojo article, SMART is just a tool, not a fail-safe system for data protection.

Proper data backup is the key to protecting your data. In the meantime, apps like DriveDX can be very useful in monitoring the health and predicting potential failure of your storage hardware.

 

[ZVH]

Gentlemen:

I am not a licensing expert or an attorney. I was under the impression that if someone used open source code based products under a GNU license, then legally their code had to be open source as well. That doesn't mean the product has to be free, but their code has to be made available for free. Like I said, I'm not a lawyer or a licensing expert, and have no interest in being one. I can afford a trial attorney in a lawsuit like I can afford to buy the Taj Mahal.

Regarding the OP, it wasn't the OP, it was the second post, so that's my error.

The reason I don't care for SMART that much is from experience. I had a 12" PowerBook with a 100G Seagate in it. It had a fairly serious head crash in the tail end of the drive. SMART utilities report that the "drive is about to die any minute now." I pulled the drive, replaced it, and then put it in a Titanium clunker figuring it would give up the ghost in short time. The Titanium is used by my little kids to play some games on. It's still working. I believe its been three years.

I had another drive that would periodically emit a squeal. The squeals were brief in duration and infrequent. I checked the SMART status several times. No problems whatsoever reported. I thought maybe it was just a fan or the optical drive making noise, after all, SMART reported nothing.

Finally, I turned on the system one day, the squeal kicked in but it didn't stop, the system locked up, and then on reboot the drive wasn't seen. The drive was dead.

That's my experience with SMART, and why I don't trust it.

 

[SaSaSushi]

Gentlemen:

I am not a licensing expert or an attorney. I was under the impression that if someone used open source code based products under a GNU license, then legally their code had to be open source as well.

As I mentioned, all of the OS X SMART apps use smartmontool to some extent to obtain SMART data. What's the deal with singling out DriveDX?

The reason I don't care for SMART that much is from experience. I had a 12" PowerBook with a 100G Seagate in it. It had a fairly serious head crash in the tail end of the drive. SMART utilities report that the "drive is about to die any minute now."

See my own experiences with SMART reporting above. I've yet to see SMART report serious errors on a drive that wasn't actually failing.

I pulled the drive, replaced it, and then put it in a Titanium clunker figuring it would give up the ghost in short time. The Titanium is used by my little kids to play some games on. It's still working. I believe its been three years.

That the drive continues to function is not proof that it doesn't have serious physical issues. SMART is only a tool. If that's all you're using that drive for then no worries but you wouldn't want to use it to store critical data, photos or for system backups.

I had another drive that would periodically emit a squeal. The squeals were brief in duration and infrequent. I checked the SMART status several times. No problems whatsoever reported. I thought maybe it was just a fan or the optical drive making noise, after all, SMART reported nothing.

If a hard disk is making abnormal sounds such as squealing, clicking or crunching you don't need SMART data to tell you something is wrong with it and that it should be replaced as soon as possible.

 

[ZVH]

As I mentioned, all of the OS X SMART apps use smartmontool to some extent to obtain SMART data. What's the deal with singling out DriveDX?

I'm not singling out DriveDX. I know of one and only one other utility that uses smartctl as the source of its actual data. Apple's smart monitoring is inherent to their system. It is not using smartmontools. There are some other tools in my list of tools and they do not use smartmontools or copies of it to do their work.

See my own experiences with SMART reporting above. I've yet to see SMART report serious errors on a drive that wasn't actually failing.

All you're doing is proving how inconsistent it is. Just like the Google article states.

That the drive continues to function is not proof that it doesn't have serious physical issues. SMART is only a tool. If that's all you're using that drive for then no worries but you wouldn't want to use it to store critical data, photos or for system backups.

That's quite correct. The issue, however, is that SMART monitoring is continuing to report the drive as "about to fail" after years of use. The drive isn't about to fail, it had a single, one-time failure. There are no more sectors being re-located. I made note of the data over time. The reason it was put in the kids toy system was because I don't trust it, but it still works. If SMART was smart, it would, over time, have recognized this as a one time event, but it hasn't.

If a hard disk is making abnormal sounds such as squealing, clicking or crunching you don't need SMART data to tell you something is wrong with it and that it should be replaced as soon as possible.

Those are the signs of serious errors, and yet SMART said nothing. Nothing!

It's just not cracked up to be what you guys are implying it is.

 

[SaSaSushi]

It's just not cracked up to be what you guys are implying it is.

The bottom line is:

SMART monitoring is the best tool we've got at present for predicting the potential mechanical failure of and monitoring the health of hard disk hardware-wise when drives aren't making strange noises or showing other symptoms.

Does SMART accurately predict every drive hardware failure? No, and no one said it did. No technology is perfect.

SmartReporter, Smart Utility and DriveDX all use Smartmontools to access the SMART data on hard disks.

As for your personal opinions as to the non-value of SMART data, I disagree.

 

[Weaselboy]

I am not a licensing expert or an attorney. I was under the impression that if someone used open source code based products under a GNU license, then legally their code had to be open source as well. That doesn't mean the product has to be free, but their code has to be made available for free. Like I said, I'm not a lawyer or a licensing expert, and have no interest in being one. I can afford a trial attorney in a lawsuit like I can afford to buy the Taj Mahal.

That is only partly true. If you use the binary unmodified in a commercial app, you just have to acknowledge that like DriveDX has. If you use the source code of the app and modify that source code you have to make the modified source code available to the community free so everybody can benefit from your add-ons or changes, but you can still use it in your app and charge for it.

OS X itself uses tons of open source code (list here) and until recently with Mavericks they charged for OS X.

There is absolutely nothing wrong with what DriveDX is doing here.

It's just not cracked up to be what you guys are implying it is.

I don't how many times you would like me to say it. I completely agree SMART is imperfect. But it just another tool in the toolbox to try and predict drive failure and IMO is of value even if it only works half the time.

 

[iMacC2D]

Removed.

 

[OldGuyTom]

@monokakata:

Don't be too alarmed by a ratio RPI, it's the linear RPI you need to be worried about. Scannerz picks up everything, including things like cable problems. It's measuring the timing. After doing a test on a drive, if any errors or irregularities are detected, select the drive and do a re-evaluation of the drive in diagnostics mode. Odds are it's just a false irregularity, but it might be a developing problem. All that stuff is in the manual.

FWIW, I bought WD external and it came with just about the cheapest USB cable I've ever seen in my life. The drive kept locking up and giving me the spinning beach balls. I thought for sure the drive was bad but it was out of the 30 day store warranty and I didn't want to be bothered with doing an RMA, so I got Scannerz because it had a good reputation and it was relatively cheap. Normally Scannerz is really boring, but when I started moving the cheap USB cable around, it went ballistic.

I thought for sure it was going to be the drive but it was the cable. You would think manufacturers would know better. I mean seriously, hard drives are complicated devices and I just figured that in terms of luck, my number had come up, but instead it was the cheapest part in the kit. The cable is THE device that sends data to the HD. Why be cheap on that, of all things.

I love Scannerz. If you have problems or questions, call or e-mail their tech support. They've been fantastic.

 

[monokakata]

Thanks. Re-running Scannerz on that drive was on my list for the day but I didn't get to it. I'll be interested to see the report.

 

[shiekh]

You really should reread the Google study. I agree SMART is imperfect, but the Google study showed 60% of the drives that failed were showing SMART errors.

So 40% failed without showing any error; so one might catch about half of the failures before they happen; good but not great.

 

[Weaselboy]

So 40% failed without showing any error; so one might catch about half of the failures before they happen; good but not great.

Better than nothing I suppose.

 

[iRock1]

Thanks for all the discussion, people.

Truth is I don't have any app at this moment, so I guess I'll have to make a call, unless I'm grabbing only free software.

I get that SMART is not perfect and works only 6 out of 10 times. However, and from what I've read in this same thread, isn't Scannerz better than that?

 

[Weaselboy]

Thanks for all the discussion, people.

Truth is I don't have any app at this moment, so I guess I'll have to make a call, unless I'm grabbing only free software.

I get that SMART is not perfect and works only 6 out of 10 times. However, and from what I've read in this same thread, isn't Scannerz better than that?

Hypothetically, SMART attributes will be reported the same in any app that reads those attributes from the drive. As far as the full disk scanning abilities of Scannerz vs. DriveDX, I have no clue which is better as I have never seen any authoritative test comparing this feature of the two apps.

 

[OldGuyTom]

Scannerz isn't just drive testing software - that's the difference.

When I had problems with the cable in the elfen-made WD I bought, I got Scannerz because it had more features, not just SMART testing. Scannerz does a scan and measures for scan errors and timing problems too, from which it can derive other info. In my case it was a bad cable.

When I got Scannerz it was right before they released the newer version with automatic diagnostics mode. THe older version wasn't hard to use, just "manual." A few months ago when they came out with the version that does automatic diagnostics, it just became that much easier to use because it can pretty much tell you if the drive has bad sectors, bad cables, and some other errors.

I generally have Performance Probe running on my Mav system all the time. PP is another tool that comes with Scannerz. I like the way it presents memory in the old way activity monitor used to. It's always entertaining watching Safari and Safari Web Content gobble up gobble up nearly 2G of memory when hitting some web sites with trivial video.