Why does the Disk Management snap-in report my volume as Healthy when the drive is dying?

Windows Vista displays a big scary dialog when the hard drive's on-board circuitry reports that the hardware is starting to fail. Yet if you go to the Disk Management snap-in, it reports that the drive is Healthy. What's up with that?

The Disk Management snap-in is interested in the logical structure of the drive. Is the partition table consistent? Is there enough information in the volume to allow the operating system to mount it? It doesn't know about the drive's physical condition. In other words, "As far as the Disk Management snap-in is concerned, the drive is healthy."

Similarly, your car's on-board GPS may tell you that you are on track for a 6pm arrival at your destination, unaware that you have an oil leak that is going to force you to the side of the road sooner or later. All the GPS cares about is that the car is travelling along the correct road.

[Raymond is currently away; this message was pre-recorded.]

Comments (41)
  1. Zem says:

    So Vista has smart support inbuild? that is good to know.

    The information in the disk management snap-in always confused me.

    The length of the ui rectangle corresponding to a partition or disk seems to be somehow related to the size of the partition or disk, but in an entriely non-obvious way. Bigger size means longer rectangle, but it is totally non-proportional.

    In some cases even a difference of a few mb is visible, but in others a 900 GB difference barely affects the size between two partitions on the same disk.

  2. whatever says:

    Since I don’t use Vista I have no clue what I am talking about, but you don’t expect that to stop me, do you? No.

    This seems like it could be really confusing. Couldn’t the SMART report simply be shown in the DM snap-in somehow if it indicated an error? I mean, first my computer tells me one thing then it tells me something completely different. Yo wonder that I, as a user, am confused, and then angry when my disc goes "pfft"?

  3. Jeff R says:

    This is poor usability.  Users don’t expect to go to the GPS to find out whether the oil is low.  Users DEFINITELY expect to use the disk management snap-in for everything related to the disk (formatting, mounting, defragging, … knowing if it’s reporting it’s about to die).

    This is Microsoft’s big problem in a nutshell…their senior technical folks offer all these elaborate explanations for why usability sucks, instead of identifying the suckage and making it not suck.

  4. Leo Davidson says:

    Might save some people some time if they go looking for SMART monitoring tools as a result of this post:

    If you’re using Vista and your motherboard is from the popular brand beginning with ‘N’ then you won’t get SMART monitoring information because their SATA drivers block the information. :(

    I tried various SMART tools and none of them could read anything. After much faffing about  I discovered some posts explaining that drivers prevent the SMART data from being radable. Fantastic, eh? How could they overlook such a thing?

    (The XP SATA drivers for the same brand are okay. For whatever reason they ship very different drivers for the two platforms. SATA hot-swap is another feature of these drivers that only works in the XP version.)

    (At least, it was this way the last time I checked, about a year and a half after Vista’s retail release. Another problem they have is it’s difficult to find release notes to see what’s changed.)

  5. Leo Davidson says:

    Zem, If you think about the range of partition sizes which the disk management snap-in has to display, would you actually want proportional sizing? You’d end up with partitions that were only a few pixels wide, if that.

  6. @Zem: the default scale of the partition sizes is logarithmic. You can change this to linear scale using the "Settings" button on the toolbar, but in this case very small partitions will be nearly invisible.

  7. James Schend says:

    Jeff, it’s much more likely that everybody realizes it sucks (now) but at the time they were developing it, the team that worked on the Disk Management snap-in had absolutely no clue anybody else was working on SMART status notifications, and so they never had a chance to really communicate cross-team and integrate the two features.

    I agree it’s not great usability, but don’t assume things are that way due to malice. There’s a saying I’ve read somewhere before which states that the design of a particular large software project starts looking like the org chart of the company that made it.

    Personally, I think this story is a great example of the flaw in test-based design: there’s no possible way to write a unit test for something like "the user sees a dialog in one place but that information isn’t propagated to the most obvious place for it". It can’t be done; you need smart, human testers with the ability to notice problems like this and the power to go cross-team and get them taken care of.

  8. Jeff R says:


    I agree, of course it’s not due to malice (wow…malice?  really?), but this is why companies have experienced folks…to identify these problems in the design before all the newer programmers write great code that has poor global usability.  That’s why Raymond and people like him get the big bux, right?

    This is definitely not a flaw in test-based design.  This is just a flaw in design.  By the time something like this gets to a test team, it’s already too late.  It should have just been designed right in the first place.

  9. Niels says:


    No, Vista (or any other version of Windows) doesn’t read SMART data, which is exactly what Raymond is pointing out here. What he’s writing here is exactly "Windows does not report SMART data."

    As for the size of the partitions in the Disk Manager, some logarithmic scaling is used by default. You can configure it to use linear scaling, but if you do you’ll soon see why logarithmic is the default.

  10. ton says:


    You make an interesting point about why windows doesn’t attempt to read SMART data since some motherboard vendors write drivers that block SMART data from being read.

    I’m curious how widespread is this and is there any viable reason they would need to block access to this information?

  11. James Schend says:

    This is definitely not a flaw in test-based

    design.  This is just a flaw in design.  By

    the time something like this gets to a test

    team, it’s already too late.  It should have

    just been designed right in the first place.

    No, because testers should be involved from the first step. You send your design documents to the testers, so they can preempt situations like this. You also show them your UI mock-ups, and whatever other documents you produce in the design phase.

    If you’re not engaging your testers until the code’s already written, you’re doing it wrong. IMO, at least.

  12. If Vista doesn’t read SMART data, how the heck does it display the "big scary dialog?" Osmosis?

    Just about everything I wanted to say has already been said. There’s more to testing than running your automated tests and calling it done when all of them indicate success. There’s more to design than creating a bunch of logical test cases.

    In the end: the user doesn’t care, JUST FIX IT. An excuse (no matter how good) is not a fix.

  13. ton says:


    From looking at the Wikipedia article I found this sentence in the article:

    ‘Many motherboards will display a warning message when a disk drive is approaching failure.’

    So based on this info the "big scary dialog" is actually coming from the motherboard driver software itself and not Vista.

  14. Bryan says:


    What do you do if SMART isn’t enabled or available or is just "wrong" for whatever reason?

    I think the assumption many people are making begs that no one asked the question "What about SMART data?" when designing the view.  What if we take it from the perspective that the question was asked (and assumed it wasn’t thrown out because of time)?

  15. manicmarc says:

    How can I tell if my laptop has SMART support? Since my 3 year old Compaq desktop has it, I assumed my 6 month old HP would – but now I have doubts!

    Is there a way to find out in Vista?

  16. DWalker59 says:

    Raymond, your term "Similarly" is off the mark.  I don’t expect my GPS to track my engine’s oil leak, but if Disk Management reports that a drive is Healthy, I would expect the drive to be Healthy.

    And I HAVE seen a drive show up in Disk Management in an "Error" state, where I thought the state was the result of I/O errors while writing to the drive.

    The term HEALTHY is a bad choice of words in this context, and it’s certainly misleading.  A better term would be "Ready" or "Logically Consistent".

    @ton:  That warning from the motherboard (from the motherboard’s IDE or SATA or SCSI driver in the BIOS) would be IN ADDITION to the one that Vista gives.  Just because a motherboard MIGHT be able to give this info, doesn’t mean that Vista can’t give its own info in a big scary dialog also.  I am not sure.

  17. Aaron G says:

    I love how so many people think that they know exactly what it takes to fix every (perceived) design flaw put out by Microsoft.  Namely, vapid one-liners that may or may not actually be true.  Folks, why don’t you apply those mad skills by working for them instead of ranting on somebody else’s blog?

  18. ton says:

    Hmm…it looks like there are ways to get to the SMART data in at least some cases. I found this doing some more research:


    Also there was this quote from the article:

    ‘The remaining part of the smartmontools package is the smartd dæmon that does regular monitoring for you. It monitors the disk’s SMART data for signs of problems.’

    It looks like its possible to get to the data its just that it may have been deemed too much trouble to implement for Vista.

  19. John says:


    Either you made a typo or you are using a definition of "popular" for which I am unfamiliar with.  Perhaps you meant ‘M’?

  20. Is there a screenshot of this big scary dialog in Vista? I want to see its UI.

  21. There’s no status indicator in my GPS that tells me the car is healthy. If there were, and it told me the car was fine while the "Check Engine" indicator was lit on my dash, I’d be a bit distressed, especially considering that I know less about the internals of my car than the average user knows about his computer.


  22. Steve D says:

    It sounds like the integrity of the drive hardware has little relationship to the inegrity of the data it contains, until… :-(

  23. Aaargh! says:

    The issue is not that DM should report the SMART status in that spot, it’s not actually that bad if you know what it means. Changing "Healty" into "Filesystem healthy" would have fixed the confusion.

  24. SRS says:

    @Jeff R – you are so correct. This is dreadful usability, and a hopeless analogy to excuse it with.

    If your disk is wrong, and the hardware is letting Vista know this through a well-known channel, then any Vista component stating otherwise is worthless – does no-one at Microsoft see the absurdity of one Microsoft supplied disk status related component saying ‘all OK’, whilst another is warning of the imminent end of the world for your data?

  25. John says:

    For some reason this reminds me of the entry a few days ago about GetActiveWindow() vs GetForegroundWindow().

  26. Steve says:

    Everyone’s getting their panties in a knot, but I didn’t read anything Raymond said that this was a good situation. So ripping on him and Microsoft more than a little bit is a bad thing.

  27. JamesW says:

    OS X shows how it should be done:


    The Windows way that Raymond describes does seem broken. ‘The Disk Management snap-in is interested in the logical structure of the drive’. That’s great for the Disk Management snap-in, but I’m sure that 99% of the users wouldn’t know what a drive’s logical structure was, much less care about it. They want to verify their drive (what magic verification does is neither here nor there) if they have cause for concern. They would also appreciate a DANGER! DANGER! warning on the screen when they go to check the drive if failure is imminent. Pity Vista got rid of animated assistants, otherwise the Lost in Space robot could be shown throwing a fit in the corner.

  28. Dog says:

    Steve: Everyone’s getting their panties in a knot, but I didn’t read anything Raymond said that this was a good situation. So ripping on him and Microsoft more than a little bit is a bad thing.

    But Raymond didn’t say it was a bad situation either. By explaining the situation, he does however give the impression that he sees no problem with it, weather he means to or not. It is impossible to report something completely objectively, the act of reporting on it gives the appearance of support unless otherwise stated.

    Raymond, IMHO this blog would be far more interesting (and you might get rid of some of the "undesirable" comments) if you gave your opinion of things… Although I’m sure MS PR have told you not to.

  29. JM says:

    DWalker59 is on the mark. The flaw here is that Disk Management proclaims volumes "healthy". That’s a vague, feel-good term that you *would* expect to include SMART assessment of the drive the volume is on. How is a volume that’s about to be reduced to a pile of access errors "healthy"? If it called volumes "consistent" expectations would be lower.

    Or, hell, why *not* make it take SMART data into account? Separation of concerns is all well and good for designing systems, but in an interface it makes sense to have integrated information available at your fingertips. Disk Management predates integrated SMART support, which explains it, but that’s no reason not to have it *now*.

    For all those folks not running Vista, *install smartmontools for Windows now*. SMART is ubiquitous, cheap and an easy lifesaver. You’d be crazy to ignore it, even if you *are* one of those smart people who make full backups every day (hands up everyone who doesn’t).

  30. Neil says:

    Nitpickers’ corner:

    Disk Management manages partitions, not disks.

  31. ton says:


    The partition wouldn’t exist without the disk would it? Also the name of the utility is "Disk Management". So either way you slice it something is wrong here…

  32. mikeb says:

    @Dog:  Raymond is definitely putting a spin on this that indicates it’s not a real problem.  Though his analogy seems a bit stretched (ie., you don’t expect your GPS to factor in all possible problems getting to your destination, why should you expect the Disk Management utility to factor in all possible problems with your disk?)

    This is another instance of the "You’re in an airplane" joke (http://blogs.msdn.com/oldnewthing/archive/2004/02/17/74812.aspx#75569).

  33. Reginald Wellington III. says:

    "(ie., you don’t expect your GPS to factor in all possible problems getting to your destination, why should you expect the Disk Management utility to factor in all possible problems with your disk?)"

    Raymond didn’t tell you what you should or should not expect from your GPS.  That’s your spin.  If you have a GPS console in your car, you may indeed expect it to tell you about an oil leak, and would be quite disappointed if you still had to rely on the "check engine" light.

  34. Peter T says:

    Not sure if Leo will see this, but take a look at http://crystalmark.info/software/CrystalDiskInfo/index-e.html for a tool that DOES support SMART data access even with drivers from the N company on Vista

  35. mikeb says:

    @Reginald Wellington III:

    I may be wrong, but I think a reasonable interpretation of Raymond’s GPS analogy is that it’s about as valid to expect the Disk Manager to take into account the physical state of the disk drive as it is to expect the GPS to factor in the physical state of your vehicle.

    I also think that most reasonable people would see that as a stretch.

  36. Leo Davidson says:

    @Peter T: Thanks! I tried it and it seems to work.

    For the people who were wondering, the problem with this manufacturer’s Vista drivers is they map SATA drives as SCSI devices rather than IDE, and SCSI apparently doesn’t support SMART. That’s my understanding, anyway.

    The tool that Peter found seems to work around the problem, though. (I’ve tried about 5 others and none worked and seen plenty of forum threads explaining what the problem was, discussing installing the XP drivers on Vista, etc.) Great to be able to see the data…

    …and now I see one of my drives has a "Caution" indicator. Uh-oh! :)

  37. mikeb says:


    As unwelcome as it might be, it’s better then getting a "Disk not found" error…

  38. steveg says:

    You can’t trust SMART data, so what’s the point? Different disk manufacturers implement different parts of it, some (apparently, I read it on the internet somewhere so it must be true) fudge it to make their drives look like they’re performing okay.

    I agree with the feeling, though, that it’s poor overall design in the Windows UI — it might be good design at the component level, but not from a user’s POV.

  39. Michiel says:

    A portable GPS isn’t a car device. It tells me about the road. And I hope it knows about roads that are closed. My car can tell me about the oil.

    Microsoft can’t take a similar approach here. For starters, they’re not two vendors. And I won’t complain that Office saves its files to an unhealthy disk. But if you can push integration to the point of stuffing a webbrowser in the OS, you should be able to integrate the management of disks into Disk Management (!)

    It might be wise to consider licensing ZFS. It takes integration of the different layers far beyond the level we just witnessed in Windows, and Sun isn’t in the best of negotiating positions today.

  40. !SMART says:

    Disk Management reads Stupid data, not Smart.

  41. So bear in mind that I’m coming at this from having spent more than ten years working for hard disk manufacturers…

    SMART is a standard the same way that OBD automotive diagnostics are, in that the spec defines an extremely minimal set of required behaviours which tell you pretty much nothing more than "go/no-go".  In order to get any of the information that you actually care about, you’re off in non-standard (and frequently undocumented) vendor-specific – and often model-specific – extensions.

    To extend the analogy futher… on top of that, the physical, protocol, and driver layers are also subject to a lot of vagaries (i.e., Volkswagen’s OBD ports stream a continuous fixed pattern and toggle the *baud rate* to signify a 1 or 0!) so you frequently need to jump through hugely different hoops to get the data you want depending on the exact combination of controller, interface type, OS, etc., etc.

    In other words, it’s freakin’ amazing to me that anyone has ever done anything useful with it.

Comments are closed.