How do you know when a HDD is going to die?

Discussion in 'Storage & Backup' started by Philll, Jul 21, 2010.

  1. Philll

    Philll (Taking a Break)

    Joined:
    Dec 30, 2008
    Messages:
    10,920
    Location:
    NSW
    I've had my current HDD's since 2007 (never really needed an upgrade) and lately they have started to 'think more' before doing anything. Is this a sign they're on the way out or that they just need a good defrag? It's especially noticeable at startup now where it wasn't so much before. This is a 2 week old Win7 install if that helps.

    Thanks
     
    Last edited by a moderator: Jul 28, 2010
  2. g3monster

    g3monster Member

    Joined:
    Feb 27, 2006
    Messages:
    732
    Location:
    Melbourne
    In my experience with dieing hard drives some signs of a hdd going bad are:

    Slower than usual seek times.
    IO rates dropping.
    Squealing when reading or writing.
    More bearing/motor noise than usual.
    Drive getting hotter than usual.
    Smart warnings.
    CRC errors.
    Clicking drives.

    Do the obvious things first replace data cable, check bios settings, check smart status for any warning signs, run a hdd benchmark program and check seek rates and speeds against specs (run this a few times and check for realitively consistant results...This works best when the hdd is not the boot drive), run chkdsk (alternatively you can use a third party tool with visual scaning to see where any bad block are) scanning for bad sectors(you can do this a few times aswell if your paranoid), defrag.

    Might also be worth running a memtest just to be safe.

    If that doesn't bring up results/fix your slowness try doing a full wipe zeroing all data (not a format) and reinstall windows

    If none of these work it's probs time for a new drive.
     
  3. flain

    flain Member

    Joined:
    Oct 5, 2005
    Messages:
    2,310
    Location:
    Sydney
    Above post is true, but i'll add the obvious bad sectors to the list

    Also "freezes" on reads or writes. Slow transfers (like 4MB/sec) etc mean its probably due for replacement
     
  4. memnoch

    memnoch Member

    Joined:
    Jan 4, 2002
    Messages:
    517
    Location:
    Sydnet
    Have you enabled SMART in the bios?
    If not, go do so now.

    Then go download speedfan (windows tool & free), use the SMART hard disk tests packaged into that tool. SMART is pretty useful to help identify drives that are likely to fail.

    Things to look for are bad blocks being re-assigned etc. If that is happening, go replace the drive.

    It is far easier to replace a drive that you can read from than restoring from backups/rebuilding failed arrays due to lack of pro-active maintenance.
     
  5. peter10001

    peter10001 Member

    Joined:
    May 28, 2010
    Messages:
    149
    Location:
    Netherlands
    -In bios enable smart.
    -Make your pc really hot, cpu burn and make your gpu hot and with hd tune do a surface scan. then see if you hd get hotter then 50 degrees, if hotter then you need more fans to cool.
    -Every 3 months do a surface scan with hd tune(you can adjust on how often you use your pc).
    -every 3 months check de smart with hd tune, especially the "reallocated sectors", if this starts growing then check more often, if it gkeeps on getting bigger then remove disk. (Maybe it is possible to put this on de background on your screen, then you see it wen you boot).
    -check in smart how often the hd goes to sleep(head goes to sleep), some new "green" harddisk go to often to sleep, and your harddisk wil break in a short time, maybe it depends on the operating system you have.

    Always keep a backup.
    Harddisk break is a second.(often i hear it is new hardware and it does not break).
    A lot of harddisks are delivered with bad sectors, so check new drives(yes it takes a long time with 2000Mb drives).
     
    Last edited: Jul 21, 2010
  6. Dezza Bot

    Dezza Bot Member

    Joined:
    Aug 16, 2004
    Messages:
    4,609
    Location:
    Central Coast, NSW
    Realistically, you don't. So you backup.
     
  7. 0shooter0

    0shooter0 Member

    Joined:
    Jan 24, 2002
    Messages:
    92
    your ibm serveRAID advises you via a automatic alert that the sas 15K has detected a PFA (predetermined failure alert), you then contact ibm support and a new one is on way way within 4 hours :)

    ohh wait.. home user? no idea... just backup.
     
  8. Aetherone

    Aetherone Member

    Joined:
    Jan 15, 2002
    Messages:
    8,543
    Location:
    Adelaide, SA
    In the spirit of fairness, most of us are in Australia. 50 degrees can be only a few degrees above summer ambient. :wired:
     
  9. pksw

    pksw Member

    Joined:
    Aug 11, 2001
    Messages:
    759
    Location:
    Melbourne
    Is there somewhere I can read about what the SMART parameters mean, on a practical level? The wiki gives some help, but are there any recommendations on what is unacceptable, even on a working drive?

    I have Everest, and one of my drives has a little exclamation mark next to it:

    ID Attribute Description Threshold Value Worst Data Status
    05 Reallocated Sector Count 36 100 100 1 OK: Value is normal

    whereas the "normal" drives show up with a combination of the following:

    ID Attribute Description Threshold Value Worst Data Status
    05 Reallocated Sector Count 36 100 100 0 OK: Value is normal
    05 Reallocated Sector Count 140 200 200 0 OK: Value is normal

    The only difference I can see is in the "Data" flag - not sure what this means.
     
  10. memnoch

    memnoch Member

    Joined:
    Jan 4, 2002
    Messages:
    517
    Location:
    Sydnet
  11. rowan194

    rowan194 Member

    Joined:
    Jan 5, 2009
    Messages:
    2,030
    That means you have 1 reallocated sector. The drive has determined it's gone bad so it has remapped that sector to a spare one. This should be considered a "caution" as it may be a sign of things to come.

    With regards to your original question, there is no way to know. SMART may be able to give you a warning but it's certainly not magic: your drive could throw a bearing, or a head could hit the platter, or an electronic part on the PCB could fail... none of these can be predicted. Backing up is the only way to keep your data safe.
     
  12. Aetherone

    Aetherone Member

    Joined:
    Jan 15, 2002
    Messages:
    8,543
    Location:
    Adelaide, SA
    I'll echo this sentiment. I've seen drives with BIOS-halt reporting smart errors continue to work fine for many months and others utterly die without any warning whatsoever.

    That said, if any of my drives chuck a smart error, I replace them post haste.

    Philll: have you actually run manufacturers diagnostics over the drives to see what they say?
     
  13. OP
    OP
    Philll

    Philll (Taking a Break)

    Joined:
    Dec 30, 2008
    Messages:
    10,920
    Location:
    NSW
    Not just yet, I'll have to do it tomorrow after work. It seems to be getting worse :/
     
  14. Hive

    Hive Member

    Joined:
    Jul 8, 2010
    Messages:
    5,177
    Location:
    ( ͡° ͜ʖ ͡°)
    That's like asking when the world going to end

    It could be 5 seconds from now or 1930983445098e+029 years from now.

    But as others have posted, there can sometimes be warning signs of failure
     
  15. Foliage

    Foliage Member

    Joined:
    Jan 22, 2002
    Messages:
    32,060
    Location:
    Sleepwithyourdadelaide
    If its in bad enough condition for you to notice it, then yes it is going to fail soon almost certainly.

    Do as suggested and before anything, back your data up, then run smart and diagnostic tools on it as some of the diagnostic tools stress test it which if its about to fail, could push it over the edge.
     
  16. Aetherone

    Aetherone Member

    Joined:
    Jan 15, 2002
    Messages:
    8,543
    Location:
    Adelaide, SA
    Echoing Foliage: Back up first.

    During my time as a tech, I've had many questionable drives fail while being tested. Diagnostics tests worth running will stress a drive - sometimes into complete failure.
     
  17. Figaro

    Figaro New Member

    Joined:
    Jul 23, 2010
    Messages:
    1
    Things are bad when :

    * Your PC will not start, black/blue screen appears with error text referring to hard disk
    * Starting your PC takes longer than normal, when using the computer downtimes occur
    * Your computer starts but you cannot open all the files you wish
    * You see an error messages like read/write error, file open error, invalid CRC etc.
    * Your computer heats up near the disk region, the temperature rise is in proportion with
    problems when starting programs, opening files - so called system freezes occur
    * Hard drive makes wheezing and clicking sounds
    * You can see smoke coming from hard drive, burn marks can be seen
    * Hard drive does not make any sound at all, the engine will not start
    * You've broken the interface terminal(s) during the installation

    DO NOT:

    * Do not try to turn the computer on repeatedly
    * Do not try to "fix" the disk by beating, knocking or shaking
    * Do not try to open the disk by yourself
    * Do not use the data recovery software
    * If the device has been exposed to water, let it dry thoroughly


    Here is more info on this subject: http://www.hotfrog.com.au/Companies/ProDigi-OÜ/Data-Recovery-from-broken-devices-114507
     
  18. oohms

    oohms Member

    Joined:
    Dec 22, 2005
    Messages:
    2,192
    Location:
    Melbourne 3015
    Check for strange noises / regular clicks / spinning up and down

    Also try doing a HDtune random access read benchmark - and post up a screenshot of the resulting graph
     
  19. SeiferCQGL

    SeiferCQGL Member

    Joined:
    Mar 1, 2008
    Messages:
    494
    Location:
    Central Qld
    It hands you its Last Will and Testament.
     
  20. DavidRa

    DavidRa Member

    Joined:
    Jun 8, 2002
    Messages:
    3,031
    Location:
    NSW Central Coast
    You're less than a minute away from starting, or completing, your backup (usually the former). That or you're away from it and unable to effect any changes on the system.
     

Share This Page

Advertisement: