Western Digital hard drives are garbage

Discussion in 'Digital Photography' started by messenjah, May 19, 2005.

  1. messenjah

    Zed Pobre Guest

    My primary job is system administration, so I work with a lot of hard
    drives, and push them somewhat hard. We don't use WD, but Maxtor has
    had the lowest survival rate of any of the other hard drives that
    we've used since they merged with Quantum a few years back. After
    having two hard drives fail in less than six months, and then the two
    replacements failing in less than six months, and then the two
    replacements of *those* failing in less than six months, I finally
    took the hint and stopped buying them completely.

    Asking other admins, I found out it's not just me. Doing a bit of
    research on hard drives in that community gave me three brands that
    are supposed to be high quality under load:

    Seagate
    Hitachi
    Fujitsu

    We're currently testing Seagate 400s (in both SATA and PATA) and are
    about to start testing a batch of Hitachi (PATA only) drives, and
    (knock on wood) the Seagates haven't failed yet (going about 6 months
    so far). I've found out that Seagate drives have significantly higher
    startup current than Hitachis, though (2.8A vs 2.0A), which can cause
    them to fail in some single external enclosure units, which tend to
    have marginal power supplies. Seagates have the better warranties,
    though, which speaks well of company confidence, at least.

    I've previously had bad experiences with IBM drives, but I'm told that
    limited to a single manufacturing batch, and they're much better than
    I think. I wouldn't know, as I haven't bought those since then,
    either.
     
    Zed Pobre, May 20, 2005
    #81
    1. Advertisements

  2. messenjah

    CWatters Guest

    Yes I got two 120G WD drives that don't have a temp sensor. Both less than
    three years old.
     
    CWatters, May 20, 2005
    #82
    1. Advertisements

  3. messenjah

    Rod Speed Guest

    Fujitsu doesnt even make 3.5" IDEs anymore and their
    last, the MPG was a complete disaster failure rate wise.
     
    Rod Speed, May 20, 2005
    #83
  4. messenjah

    Zed Pobre Guest

    They still make SCSI drives, the MA* series, as I understand it. I
    never looked into Fujitsu personally, so I don't know much about them.
    They just showed up in conversations as solid drives.
     
    Zed Pobre, May 20, 2005
    #84
  5. messenjah

    JPS Guest

    In message <>,
    I didn't say otherwise, but the fact is, most people blame down-time
    spent on mechanical seeks on something else, usually the CPU. People go
    out and buy brand-new computers with lots of GHz because their systems
    are sluggish, but it isn't necessarily the CPU that is the problem; most
    often it is disk head contention.

    A drive capable of reading a single, contiguous file at 50 MB/s can drop
    4 or 5 MB/s total if two files are being accessed at different ends of
    the drive at the same time.

    --
     
    JPS, May 21, 2005
    #85
  6. messenjah

    JPS Guest

    In message <>,
    Perhaps. I said it was one explanation, as I heard this most recently,
    from someone who said I was wrong when I claimed it had something to do
    with the filament being cold.
    I've always thought it was something like that.

    --
     
    JPS, May 21, 2005
    #86
  7. messenjah

    JPS Guest

    In message <>,
    You can say anything you want about me, but that has no effect on what I
    am or what I do.
    --
     
    JPS, May 21, 2005
    #87
  8. messenjah

    JPS Guest

    In message <>,
    Of course they can't. Who said they did? Not me. I said that one
    explanation is that the bulbs suck in oxygen when they cool down. I
    didn't even say it was my explanation. The person who suggested that to
    me said that the seal can break as the gas contracts, bringing in oxygen
    which causes the filament to burn open.
    No kidding.
    --
     
    JPS, May 21, 2005
    #88
  9. messenjah

    JPS Guest

    In message <>,
    It wasn't my explanation. I said it was *one* explanation, given to me
    by someone else. The idea was that as the gas in the bulb contracted,
    air with oxygen from the environment leaks into the bulb, so the
    filament burns when it restarts.

    There is nothing in my statement to suggest either that this was my
    idea, or that the explanation was that oxygen was always in the bulb,
    and got sucked into the filament. That is due to sociopathic nature of
    most of the people who replied to me; people who look for any angle to
    ridicule someone from, with their interpretations of what they read
    twisted to make other people look stupid.
    --
     
    JPS, May 21, 2005
    #89
  10. messenjah

    ^reaper^ Guest

    -----BEGIN PGP SIGNED MESSAGE-----
    Hash: SHA1

    While sipping absinthe, heard a loud sucking noise coming
    from alt.2600, and hastily inscribed the following unintelligible
    Do any of you know at all what you're talking about? Or are you just
    blowing hot air to pass time? o_O

    ^reaper^

    -----BEGIN PGP SIGNATURE-----
    Version: PGPfreeware 6.5.8 for non-commercial use <http://www.pgp.com>

    iQA/AwUBQo5+ZlMeYoHj2dI5EQKOhgCfRmav9UXVV2YuM36uYDmniCJpjMsAniqa
    r9CVPD42SMWIbQ2aL0Ia5aqy
    =JwAJ
    -----END PGP SIGNATURE-----
     
    ^reaper^, May 21, 2005
    #90
  11. They're too chickenshit to come down here in the swamps.
    :)

    --
    --------------------------------------------------------------------------------------------------------------------
    Yomamma bin Crawdaddin www.cotse.com
    Brotherhood (Vice Chairman)
    Anti Archangel #41 The difference between John Kerry and Ted Kennedy is
    Stalking Filth #69.5 that Ted Kennedy has at least one confirmed kill.
    --
     
    Yomamma Bin Crawdaddin, May 21, 2005
    #91
  12. messenjah

    JPS Guest

    In message <e9xvk4zft9y0$>,
    I've tested this many times, and the results are consistent. You get
    about 15 to 25x the disk speed when two files being read "at the same
    time" are on different drives, than when they are on the same drive at
    different extreme cylinders.
    --
     
    JPS, May 21, 2005
    #92
  13. messenjah

    Rod Speed Guest

    No perhaps about it, completely mindless silly crap.
    It isnt. Its just plain mindless silly crap.
    Yeah, the 'suck in oxygen' is completely silly, there
    is an inert gas in most incandescents, for a reason.
     
    Rod Speed, May 21, 2005
    #93
  14. messenjah

    Ari Rankum Guest

    I know I'm too small a sample size to be statistically significant, but
    I'm still going to bitch.

    I've lost all three of the last WDs I've bought, just after the warranty
    expired. The only good thing I can say is that they failed gradually,
    and there were significant warnings in /var/log/messages prior to the
    failure. This allowed me to migrate to new drives relatively smoothly.

    The most recent purchase, based on experience with WD, was a Seagate.
     
    Ari Rankum, May 21, 2005
    #94
  15. messenjah

    ^reaper^ Guest

    -----BEGIN PGP SIGNED MESSAGE-----
    Hash: SHA1

    While sipping absinthe, heard a loud sucking noise coming
    from alt.2600, and hastily inscribed the following unintelligible
    For starters, if you are going to use such an example to back your
    claim, you need to qualify "tested" and "many times" as well as
    clarify under what conditions you ran your "tests," how you were able
    to determine on what cylinders these files "phyisically" resided, and
    how you were able to control your concurrent reads. You also need to
    describe how you identified and accounted for confounds when you ran
    your "tests" as well as how you established your normative values for
    later comparative analysis.

    Furthermore, do you really mean disk speed or are you referring to
    transfer rate or seek times? If the latter, how did you measure seek
    times and what do you think accounts for the extreme variance? And
    yes, an order of magnitude per second is extreme variance when you
    are talking about seek times. Especially given that the example you
    have provided involves concurrently reading two contiguous files, one
    on the low cylinder and one on the high. In fact, even if you're
    referring to transfer rate, you are still talking about a large
    variance that cannot be accounted for by seek time alone.

    And finally, while seek times can impact overall processing time, you
    seem to completely ignore the primary problem. Computes and page
    swapping. And this is driven in part by processor design and speed,
    memory management, how much resident memory is available at any given
    point in time, and of course, the number and type of program(s)
    running. Obviously if you're doing intensive disk i/o processing, you
    will see an impact from your seek times. Which is why, for example,
    large database designs include both a logical and physical design.
    Even so, to make a generalized statement that, and I quote, "HD seek
    times are the biggest bottleneck in most computing" is... well...
    completely imbecilic.

    ^reaper^

    -----BEGIN PGP SIGNATURE-----
    Version: PGPfreeware 6.5.8 for non-commercial use <http://www.pgp.com>

    iQA/AwUBQo6uxlMeYoHj2dI5EQLdHwCeKFnoCMD8H/9xaZQeFitBssR6L5wAnifw
    SBL50SePLRHjuR52btz1s5lB
    =ASmf
    -----END PGP SIGNATURE-----
     
    ^reaper^, May 21, 2005
    #95
  16. The Seagate is probably going to be the best bet. I run a pile of
    Maxtors without any hiccups, but that doesn't really mean anything. Any
    higher end builds I do, or any Servers I put into the mix that need to
    do anything more than share out a printer will receive a Seagate drive.
    The hitachis are noisy but fast, and from working with alot of Touch
    machines, I know first hand how bad the failure rate is on the Fujitsu
    drives, so I have steered clear of them for about 4 years now.

    Wheats
     
    Wheat Muncher, May 21, 2005
    #96
  17. Sounds like a troll to me. Why waste time on such useless drivel.
     
    Bullwinkle. J. Moose, May 21, 2005
    #97
  18. messenjah

    Me Guest

    I've been in those swamps... I understand why. ;-)
     
    Me, May 21, 2005
    #98
  19. Me too....
    It makes me want to grab their ass with both hands.
    :)
    --
    --------------------------------------------------------------------------------------------------------------------
    Yomamma bin Crawdaddin www.cotse.com
    Brotherhood (Vice Chairman)
    Anti Archangel #41 The difference between John Kerry and Ted Kennedy is
    Stalking Filth #69.5 that Ted Kennedy has at least one confirmed kill.
    --
     
    Yomamma Bin Crawdaddin, May 21, 2005
    #99
  20. What?
    You dissin' my home?
    You know,
    the next time I make it to Eufaula,
    I might just cross the river and come over there
    and teach you some respect !!!!!!
    :)
    --
    --------------------------------------------------------------------------------------------------------------------
    Yomamma bin Crawdaddin www.cotse.com
    Brotherhood (Vice Chairman)
    Anti Archangel #41 The difference between John Kerry and Ted Kennedy is
    Stalking Filth #69.5 that Ted Kennedy has at least one confirmed kill.
    --
     
    Yomamma Bin Crawdaddin, May 21, 2005
    1. Advertisements

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.