Submit Hint Search The Forums LinksStatsPollsHeadlinesRSS
14,000 hints and counting!

Disk Utility may not warn of a failing hard drive Storage Devices
While Apple's bundled Disk Utility application reports a hard drive's S.M.A.R.T. status, the information might not be correct.

If you care about the information stored on a hard drive that is connected to an internal bus of your Mac, you should consider using other applications than Apple's Disk Utility to monitor that drive's health.

As can be seen in the following screen shot an HDD riddled with bad blocks is still reported as having it's S.M.A.R.T. status verified.

Here's a screenshot showing the difference between Disk Utility and SMART Utility at displaying the S.M.A.R.T. parameters.

[crarko adds: S.M.A.R.T. in general has not been incredibly successful at forecasting drive failure. Here's the report from a study Google conducted on predicting drive failures and the conclusion is that some SMART parameters are more useful than others, but "Given the lack of occurrence of predictive SMART signals on a large fraction of failed drives, it is unlikely that an accurate predictive failure model can be built based on these signals alone."

There's still no better defense against drive failure than known good backups. RAID 6 is nice too, but not found on too many desktops.]
    •    
  • Currently 3.50 / 5
  You rated: 4 / 5 (6 votes cast)
 
[14,105 views]  

Disk Utility may not warn of a failing hard drive | 16 comments | Create New Account
Click here to return to the 'Disk Utility may not warn of a failing hard drive' hint
The following comments are owned by whoever posted them. This site is not responsible for what they say.
Disk Utility may not warn of a failing hard drive
Authored by: buz on Sep 30, '10 07:53:39AM

Read the logs. If I remember smartd writes about failing disk or problems detected in the system.log. That's how I saved my data ac ouple of days before the failure.



[ Reply to This | # ]
Disk Utility may not warn of a failing hard drive
Authored by: BMarsh on Sep 30, '10 08:48:28AM

SMARTreporter not only checks SMART - allows you to run SMART tests (beyond the normal SMART status), but it now also checks for I/O Errors in the system log.
And it is free.

(I have it set for close friends & family to email me if it runs across an error, in addition to putting up a dialogue on the system to warn the user)



[ Reply to This | # ]
Disk Utility may not warn of a failing hard drive
Authored by: Reaperducer on Sep 30, '10 09:54:23AM

I just looked at Smart Utility linked to in the tip, and Smart Reporter (Google it), which was mentioned in the comments.

Smart Utility: $25 for one computer, $40 for multiple computers, $100 for a business.

Smart Reporter: Free, open source (MIT).

Smart Utility seems to have a lot more information and options, but not all of them are necessarily useful for the average person. Smart Reporter seems to have all the basics, plus a few extra goodies. It's kind of "Apple" in the sense that it's powerful enough to get the job done, but not so cluttered with seldom-used options that it starts looking like Windows.



[ Reply to This | # ]
Disk Utility may not warn of a failing hard drive
Authored by: JackSloth on Sep 30, '10 12:13:15PM
Backup Backup Backup!

Absolutely backups!

Having noticed a trend of increasing amount of failing drives over the last 5 years, backup is the only sure fire way to protect yourself.

Never rely on S.M.A.R.T. status or any other utility.
---
Mac computer tech support in Montreal


[ Reply to This | # ]
Disk Utility may not warn of a failing hard drive
Authored by: adrian.nier on Sep 30, '10 01:12:26PM

Every user, average or pro, must know how many bad blocks their hard drive contains. The drive should be replaced as soon as it develops the first bad block. Otherwise, what good is a backup, if the data on your primary storage device is getting corrupted?

The gist of this hint is: Disk Utility ignores bad blocks and will not report the drive as failing, if it has any. Get yourself better software to monitor your hard drive.

There are different kinds of utilities that do so, I never mentioned, let alone link to, SMART Utility in my original hint submission except showing it in the screenshot and Iím not affiliated with its makers in any way.



[ Reply to This | # ]
Disk Utility may not warn of a failing hard drive
Authored by: gabester on Sep 30, '10 01:34:39PM

I wouldn't go so far as to say that a drive needs to be replaced as soon as the first bad block developed, but it certainly needs to considered unreliable for permanent data storage. And Mac OS X really doesn't do a very good job of informing the user about bad blocks and how to resolve them. In my mind, when a bad block is detected the OS should flag it and never attempt to use it again; in practice OS X will continually get hung up on that spot...

I've used tips from this page: http://uncoy.com/2006/09/bad_sector_io_e.html as it has good guidance on how to actually get around the bad blocks on your drive. It can be summarized as follows:
boot from OS X disc

% dd if=/dev/disk3 of=/dev/null conv=noerror

count number of errors
back up critical data
erase entire disk (select drive not partitions) - write zeros (once is enough)
then rerun:

% dd if=/dev/disk3 of=/dev/null conv=noerror

Following those steps should return zero errors. if it still errors drive is no longer reliable, check your warranty.

[ Reply to This | # ]

Disk Utility may not warn of a failing hard drive
Authored by: adrian.nier on Sep 30, '10 03:36:13PM

While the procedure is perfectly sound, it should be neither the userís nor the system administratorís job to fix something that shouldnít be broken in the first place. Most hard drives come with a 3-year, some even with a 5-year warranty. There is no bad blocks policy like the dead pixel policies some display manufacturers have. Demand a quality product, you paid for it.



[ Reply to This | # ]
Disk Utility may not warn of a failing hard drive
Authored by: crarko on Sep 30, '10 01:41:04PM

Yeah, I added the link to SMART Utility after seeing it in the screen shot; I thought people might be curious in knowing what it was.

I'm not affiliated with them either.



[ Reply to This | # ]
Disk Utility may not warn of a failing hard drive
Authored by: fahirsch on Sep 30, '10 03:34:02PM

Whenever I have to service a mac the first thing I do is use Smart Utilities. It's fast and tells me enough to report to the owner if the drive needs replacement. Again and again I have seen Disk Utilities say there was no problem and the problem was a disk failing. Smart Utlities is fast and simple. Many times I tell a client to download it and run. If the result is Failed or Failing I tell them to go to Apple and have the disk changed
It may be overkill for the home user, but for the technician it's extremely useful.



[ Reply to This | # ]
Disk Utility may not warn of a failing hard drive
Authored by: stepwallace on Sep 30, '10 06:39:53PM

Disk Utility is correctly reporting the SMART status, it's just not reporting the number of bad blocks.

Drive manufacturers set a maximum acceptable number for each measured trait of the drive before a drive will report a SMART issue. A couple of bad blocks is considered normal over the lifetime of a drive, so it won't flag as failed. Once that count has gone over what the manufacturer has determined as the maximum acceptable limit, that's the point where the drive will report as failing in SMART status.

If you open up SMART reporter, you can see the thresholds listed, and the point at which it would fail.

---
__
http://www.step.net.au



[ Reply to This | # ]
Disk Utility may not warn of a failing hard drive
Authored by: adrian.nier on Oct 01, '10 09:00:41AM
Seems to me that some of the values displayed by SMARTReporter are long past their thresholds:

ID# ATTRIBUTE_NAME          FLAG     VALUE WORST THRESH TYPE      UPDATED  WHEN_FAILED RAW_VALUE
  1 Raw_Read_Error_Rate     0x000f   100   064   046    Pre-fail  Always       -       222065
  2 Throughput_Performance  0x0005   100   100   030    Pre-fail  Offline      -       19267584
  3 Spin_Up_Time            0x0003   100   100   025    Pre-fail  Always       -       1
  4 Start_Stop_Count        0x0032   099   099   000    Old_age   Always       -       5083
  5 Reallocated_Sector_Ct   0x0033   095   095   024    Pre-fail  Always       -       127 (1875, 125)
  7 Seek_Error_Rate         0x000f   100   100   047    Pre-fail  Always       -       3301
  8 Seek_Time_Performance   0x0005   100   100   019    Pre-fail  Offline      -       0
  9 Power_On_Hours          0x0032   085   085   000    Old_age   Always       -       7942
 10 Spin_Retry_Count        0x0013   100   100   020    Pre-fail  Always       -       0
 12 Power_Cycle_Count       0x0032   100   100   000    Old_age   Always       -       3196
192 Power-Off_Retract_Count 0x0032   100   100   000    Old_age   Always       -       201
193 Load_Cycle_Count        0x0032   080   080   000    Old_age   Always       -       410036
194 Temperature_Celsius     0x0022   100   100   000    Old_age   Always       -       26 (Lifetime Min/Max 9/48)
195 Hardware_ECC_Recovered  0x001a   100   100   000    Old_age   Always       -       301
196 Reallocated_Event_Count 0x0032   095   095   000    Old_age   Always       -       127 (14, 6827)
197 Current_Pending_Sector  0x0012   091   091   000    Old_age   Always       -       10
198 Offline_Uncorrectable   0x0010   067   067   000    Old_age   Offline      -       67
199 UDMA_CRC_Error_Count    0x003e   200   200   000    Old_age   Always       -       0
200 Multi_Zone_Error_Rate   0x000f   100   100   060    Pre-fail  Always       -       2063
203 Run_Out_Cancel          0x0002   100   100   000    Old_age   Always       -       433762336302
240 Head_Flying_Hours       0x003e   200   200   000    Old_age   Always       -       0


[ Reply to This | # ]
Disk Utility may not warn of a failing hard drive
Authored by: softraid on Sep 30, '10 09:19:36PM

There is another option when you suspect a drive may be failing, and that is the SoftRAID software. You can use it for 30 days free. If all you want to do is test a suspected bad disk, or check to see if you have any disks likely to fail, download it and try it out.

SoftRAID 4.x has many tools for checking disks.

The best is "Predictive Failure Analysis". SoftRAID engineering studied the Google findings extensively, and incorporated an analysis of the findings of this study, which allow the SoftRAID application and driver to reference patterns in SMART data to determine when a disk is "highly likely to fail" in the next few months. This only works on internals, or externals which can pass SMART data to the system. Its an incredibly powerful feature.

SoftRAID also has disk "certification", which lets you run 1 to 8 passes on a disk, which will write a pattern to the disk, then read it back. 3 passes should suffice for typical testing. At the end of the certification, a disk is tested with intensive random IO activity for 15 minutes (if this is a well cooled disk, you can extend this time, but be aware that this can damage a disk that does not have adequate cooling.) Certification will enable you to take a new disk and be highly confident of it before putting it into service. Most disks that fail in the first year, will fail in the first 30 days, and certification is one way to avoid those disks. Yes, brand new disks can fail a 3 pass certification test.

There is also verification, which can be run on either disks and volumes. (Because SoftRAID is a RAID software application, it is often important to verify that both disks in a stripe/mirror are readable, and in the case of Mirrors, that both disks are identical). Verification is a standard read of an entire disk/volume, except it can be run in the background, without affecting your use of the system.

There is error tracking if you choose to use SoftRAID to control your disks. Any disk that is controlled by SoftRAID tracks and logs all disk errors. We have found that OS X ignores most (80%+) disk errors, not letting the user know they happened. This is partly why disks fail "suddenly", as you are not getting the clues that a drive is failing until it is too late.

SoftRAID is normally used on Servers and "prosumer" systems, and these disk management capabilities are pretty unique. We are investigating whether we should integrate some of this technology into a standalone product, but tracking disk errors would require our driver to control the disk. In the meantime, feel free to download the application and use it for testing your disks.



[ Reply to This | # ]
Disk Utility may not warn of a failing hard drive
Authored by: stokessd on Oct 01, '10 05:39:50AM

Be careful, S.M.A.R.T. really only is useful for drives that are slowly and gracefully degrading. In my home life I've replaced three drives in my raid in as many years. I've got smartd set to email me if it detects anything wrong with the drives (linux box). I'll get an email that a drive is dead and I'll check the logs and it was fine until it died. Bam, just toast.

In my experience smart is really good about telling you the obvious (you don't have a drive anymore), but pretty hit or miss about warning you before the fact.

Backups are the only way to have a reasonable probability that you wont lose data. No backup means sooner or later you will lose data.

Sheldon



[ Reply to This | # ]
not so SMART...
Authored by: petersconsult on Oct 01, '10 05:15:32PM

This may seem completely off-topic, but when i worked at my University's computer store ten years ago, the official life expectancy of a 2.5in 7200RPM hard drive was two years, i gather most brands now have brought that down to the warranty period: one year.

i see a lot of drives fail (my tasks include IT work), and i can't remember a single one of those failed drives that had a negative SMART report...

To be honest, i never even thought anyone serious even paid attention to the 'SMART report'.
Cheers,
Peter



[ Reply to This | # ]
not so SMART...
Authored by: j-beda on Oct 02, '10 05:49:07AM

I have had SMARTReporter kick up a warning before a drive completely bit the dust, so at least it at least occasionally does give useful information.

The Wikipedia article http://en.wikipedia.org/wiki/S.M.A.R.T. mentions that Google did find it to have some predictive ability in their testing.



[ Reply to This | # ]
not so SMART...
Authored by: petersconsult on Oct 02, '10 06:37:53AM

Sorry, i just realized that my language was a little rough...
i guess i didn't mean that it has no predictive value at all; just that, in my experience, i never relied on it as one of my 'core' troubleshooting steps/tools...
Cheers,
Peter



[ Reply to This | # ]