31

So I have a desktop PC with an SSD and an HDD. Windows 10 is installed on the SSD, all my data is on the HDD. Today while I was playing a game, it suddenly became unresponsive. At the same time, the video I was watching on my second monitor froze as well. The game is installed on the hard drive, the video file was on that hard drive as well. Both programs became unresponsive, while all other programs that are installed on the SSD continued to run fine. After about 2 minutes, the game crashed and the window disappeared, along with VLC player (which was playing the video file). I checked the Explorer, and the E: and D: partitions on that hard drive didn't show up anymore.

So I rebooted and the drive shows up again. . I ran CHKDSK on the drive, which ran fine and returned no errors or bad sectors (CMD output below). I also installed SeaTools and ran a couple of the simple tests on the drive, all of which ran fine and didn't show any error.

So now I'm not sure what to do. All my data is continuously backed up, so if the drive crashed and didn't come back, I wouldn't lose much. Since the SSD doesn't seem to be affected, I could probably just install a new hard drive without even having to reinstall Windows. But of course I don't want to throw away a perfectly good 2 TB hard drive, and since everything seems to be fine ...

What could have caused this problem? How likely is it that it was a software defect rather than a hardware problem? I noticed earlier that the video froze a couple of times, but I was copying some files and downloading some other stuff, so that's not that unusual ... are there any other diagnostics I can perform to see if there's a problem with the hard drive? Any suggestions are welcome. Thanks!

C:\WINDOWS\system32>chkdsk D: /F
The type of the file system is NTFS.

Chkdsk cannot run because the volume is in use by another
process.  Chkdsk may run if this volume is dismounted first.
ALL OPENED HANDLES TO THIS VOLUME WOULD THEN BE INVALID.
Would you like to force a dismount on this volume? (Y/N) Y
Volume dismounted.  All opened handles to this volume are now invalid.
Volume label is Data.

Stage 1: Examining basic file system structure ...
  220416 file records processed.
File verification completed.
  360 large file records processed.
  0 bad file records processed.

Stage 2: Examining file name linkage ...
  273218 index entries processed.
Index verification completed.
  0 unindexed files scanned.
  0 unindexed files recovered to lost and found.

Stage 3: Examining security descriptors ...
Security descriptor verification completed.
  26402 data files processed.
CHKDSK is verifying Usn Journal...
  37481040 USN bytes processed.
Usn Journal verification completed.

Windows has scanned the file system and found no problems.
No further action is required.

1400318975 KB total disk space.
1195790564 KB in 192027 files.
     63504 KB in 26403 indexes.
         0 KB in bad sectors.
    365935 KB in use by the system.
     65536 KB occupied by the log file.
 204098972 KB available on disk.

      4096 bytes in each allocation unit.
 350079743 total allocation units on disk.
  51024743 allocation units available on disk.
  • 58
    “All my data is continuously backed up” FINALLY someone who does it. :D It's comforting to read your message after a lot of data-recovery questions. :P – Andrea Lazzarotto Nov 13 '17 at 1:00
  • 1
    If it's a Seagate 2TB drive, it isn't "perfectly good". The Seagate desktop drives from that era have a habit of failing frequently. – Mark Nov 13 '17 at 9:53
  • 1
    @Mark WesternDigital Green 2TB, should be fine – MoritzLost Nov 13 '17 at 9:56
  • 1
    What makes you think this problem has anything to do with the hard drive? To me it sounds like you simply overheated (or triggered a bug in) your graphics card, which is not terribly surprising considering what you were doing with it. Reboot and move on? – Lightness Races in Orbit Nov 13 '17 at 11:03
  • 4
    @LightnessRacesinOrbit Why would an error with the graphics card cause the hard drive or the mapping of partitions to drive letters to crash? – MoritzLost Nov 13 '17 at 11:22
37

If the disk only crashed once and now everything is OK, then this event was maybe exceptional.

My own guess is that the hardware overheated because you were using a lot the GPU and/or the CPU.

What you can do to check the problem :

  • Look for a useful error or warning in the Event Viewer
  • Install a product that can show the disk S.M.A.R.T. data and tell you if there is a problem (example Speccy).
  • Install a temperature-monitoring product such as Speedfan which can be configured to display the current temperatures in the taskbar for easy monitoring.

If heating up is the problem, you may :

  • Verify that the game uses the video GPU and not the CPU
  • Clean up all airways
  • If a laptop, place it at an angle, so the air can pass below it, or buy a cooling pad
  • Renew the CPU thermal paste and heat sink (better done by a professional)
  • Replace the CPU heat sink and fan with a better one (also better done by a professional)
  • 16
    It could always be Windows itself getting itself up in a knot, for example because of a resources leak until it ran out of them. Just ignore it, unless it becomes a recurring problem. – harrymc Nov 12 '17 at 20:50
  • 14
    @harrymc "Just ignore it, unless it becomes a recurring problem." That's a superb example of Windows philosophy applied that survives from version to version, along with the "did you try to reboot?" one. ;-) – Hastur Nov 13 '17 at 8:33
  • 3
    @Hastur: Quoting from your link : "This Isn’t Just About Windows". – harrymc Nov 13 '17 at 8:37
  • 1
    @MoritzLost Look back around the time that stuff stopped working in the 'System' log. Look to see if you have any warnings or errors listed around that time. Even non-error/warning informational messages around that same time may have useful information. In particular, look for things from "Disk," "Virtual Disk Service," "Logical Disk Manager," etc. – reirab Nov 13 '17 at 16:33
  • 1
    @Hastur In the good ol' DOS-based days, restarting your computer after a failure was a good idea, due to the non-existence of the noble segfault. Nowadays... it's an excuse for sloppy coding (and not fixing existing sloppy coding). – wizzwizz4 Nov 13 '17 at 16:57
10

The command you ran: chkdsk d: /f doesn’t do any checking of the drive surface.

The proper command would be: chkdsk d: /r this will perform a full surface scan. If bad blocks are discovered the drive is failing.

You can also use SeaTools and do a long test. Short tests do not check the drive surface.

Chances are the drive is failing. I’m assuming it’s a Seagate if you are using SeaTools. Seagates are not only notoriously low quality drives, they do exactly what you described when they start getting bad sectors. Instead of recovering, they shut down completely until power cycled.

Skirting the line of personal opinion here, it’s not really a loss to get rid of it now before it inevitably dies. Stay away from that brand.

  • Comments are not for extended discussion; this conversation has been moved to chat. – DavidPostill Nov 14 '17 at 22:55
  • 1
    Can you include trustworthy sources that back up your claims? – Pedro A Nov 15 '17 at 0:50
  • @Hamsterrific bioinformare.blogspot.com.au/2016/02/… take it for what you will. Unfortunately it doesn’t show the older 1tb and smaller barracudas which are far worse. – Appleoddity Nov 15 '17 at 1:24
5

Move on

In your question you seem to ask which things you 'can' do, but in the title you ask what you should do.

To list the key points:

  • You have a backup
  • The problem only occurred once
  • You did some checks, which came out clean

At this point it simply not time efficient to keep digging deeper. (Unless you enjoy it). If the problem persists, you can always decide to go deeper in your search, or replace the drive.

Of course some generic maintenance (e.g. cleaning the fans) can never hurt, but don't spend too much time on it.

  • That's a pragmatic approach, this is probably appropriate unless it becomes a recurring problem. Thanks! – MoritzLost Nov 13 '17 at 13:33
  • Second this, move on. Consumer level equipment dies, and will crash without warning. Just have schedule back-up, or take full snapshot of your hard disk drive. – Chad Nov 14 '17 at 2:34
4

If a partition "doesn't show up anymore", the first thing to check would be what diskmgmt.msc sees - is the partition itself gone (which would mean something damaged the partition table - which is a rare thing unless malice is involved), or is the filesystem in that partition damaged beyond being recognized (use filesystem specific tools), or has something simply unmapped the drive letter?

If the reliability of a drive is suspect, generic SMART utilities (eg the mentioned CrystalDiskInfo) are your friend - try to understand as much as you can of the actual SMART values displayed, not just the conclusion a tool draws from them. "Pending sectors" that aren't going away fast mean the drive should be going away fast. Overtemperature or spinup failure events also can make a drive suspect.

2

I'd suggest verifying all your HDD connections are tight as well. SATA connectors hold better than the old PATA ribbons did; but years ago I had a problem with a drive occasionally dropping out until after a restart that apparently was due to a data cable that was slightly loose. After the 3rd or 4th incident I reseated all the connections and the drive worked perfectly well for the next several years - when I retired it in favor of a larger capacity model.

1

It sounds like a software problem. Try these:

  1. Check if the drive is ok using partitioning software (MiniTool free is my favorite)

  2. Head into Disk Management and check the disk is assigned to a letter

  3. Mount the drive to another PC

  4. Try checking for an issue using recovery software (I prefer Lazesoft's free recovery software, remember recovering isn't repairing, you will need space available to copy the found files)

  5. Try checking the drive health with CrystalDisk

  6. If all else fails, all I can think of is wiping the drive.

  • Everything seems to be working now. I ran a couple of diagnostics and checked the disk management, I can't see any problems there as well ... By wipe you mean reformat the drive? I could do that, if it would help ... would lose my Ubuntu installation, but there's nothing important there, and all data is backed up... but if the drive is failing, what will that accomplish? – MoritzLost Nov 12 '17 at 20:36
  • 2
    Have you ran CrystalDisk? It will give you the S.M.A.R.T health status of the drive crystalmark.info/download/archive/CrystalDiskMark/… If it says everything is ok, a piece of software someone is messing with it. If not the disk is failing and I recommend buying a new 2TB disk and cloning the disk with something like MiniTool – Brendan Jennings Nov 12 '17 at 21:12
  • Unfortunately, it's not the case that the drive is fine because the SMART data says its fine. SMART health status is often roughly correlated with the actual health of the drive, but there's no guarantee. As an extreme example, SMART won't tell you about an impeding head crash on a rotational drive. – a CVn Nov 14 '17 at 16:21
1

As harrymc said, this is very likely to have been an overheating incident.

It has probably nothing to do with Windows, as others suggested: some computers (especially laptops, but possibly desktops too) have overheating protections that cause the southbridge/PCH, where the hard drives and other things are attached, to shut down.

Look for the THERMTRIP# or THRMTRIP# signals in the Intel's chipsets datasheets (e.g. Intel® 200 and Z370 Series Chipset Families PCH Datasheet Vol. 1 ).

Usually (well, on my laptop, at least) after a couple minutes from this the whole computer will shut down.
It might not have happened to you because you rebooted it right away.

I think in some cases I did manage to keep the pc running even after the chipset's shutdown, however, by cooling it down quickly.
Of course it's of little use in most cases, since you'll usually have no way to save the work you're doing.


Anyways, especially since you were playing, it's extremely likely that this was not a hard drive or other device's failure, but just an instance of overheating.

Get into the habit of keeping a temperature-monitoring product running, possibly with alerts set up, and see if you reach dangerous temperatures often enough to warrant a cooling system overhaul (or at least cleaning).

1

Buy a new drive, but rather than “wasting” the old one, use it as an off-site backup. You can use the packaging that the new one came in, tape it up and label the package. Leave it at a relative’s house for safe keeping.

  • Are you seriously suggesting using a potentially suspect drive to store a backup copy for use in case the working copy goes bad for some reason? That sounds rather backwards to me; I'd want the backup copy to be more reliable, especially if you aren't regularly ensuring that the backup is still good and readable. – a CVn Nov 14 '17 at 16:14
  • Don't misunderstand, @MichaelKjörling. I'm not suggesting this take the place of any existing backup. But not throwing out the drive can help him feel better about “wasting” it. – JDługosz Nov 14 '17 at 17:45
  • 1
    @JDługosz "That depends". In my case, I had a Seagate 1.5 TB that was running unusually hot for a drive. Hot enough that you wouldn't want to touch it for more than a second or two. So I replaced it with a WD 2 TB to match the other two drives in my backup plan. The Seagate got wiped, and will hit the local e-waste center next time I go there. No shame in that, if a drive is failing, it seems like a really bad idea to use it for anything. Don't stress "wasting" it. Drives are cheap. Date is valueless: I have source code on those backups for personal projects that goes back 20+ years. – dgnuff Nov 14 '17 at 23:19
1

I see some good comments already (partition analyzers, heat checks, cable checks, tools). Without more information like brand/model it will be difficult to suggest anything besides that.

Tools I also tried in the past (besides the ones already mentioned):

  • Data Lifeguard Diagnostics for Windows (I used this on some WD external drives and once found a drive that was bad)
  • HD Tune (I used this for performance testing, could be an additional thing to do next to all previous mentioned checks)
  • I've used them years ago. With the WD tool I remember to have found a disk with issues so I knew I could toss it out of the window. And with HD Tune I did some performance tests, but it was long time ago, can't remember the exact results, drive seemed to perform fine, so apparently I did not use it on a troubled drive, just thought it might be a nice addition to the mentioned solutions. – Mark Duivesteijn Nov 14 '17 at 10:09
  • I've altered the answer a bit to clearify my experience with the tools. Is that what you where looking for @Pierre.Vriens? I'm on the road now, can try to find the results in a backup at home if I still have it somewhere. – Mark Duivesteijn Nov 14 '17 at 10:27
  • A bit better already @Mark, merci for taking my advice ... – Pierre.Vriens Nov 14 '17 at 10:43

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.