About the VIA Sata controller on a K8V (non deluxe)

Discussion in 'Asus' started by Mtimerding, Oct 13, 2005.

  1. Mtimerding

    Mtimerding Guest

    I had two Maxtor 300gb drives hooked up to the
    VIA Sata Controller on my K8V motherboard ....
    not in any kind of raid config, just as two additional
    drives (Drive J and Drive K).... I was keeping my
    dvd backups on those two drives. All during the time
    I had this set up I was getting a bunch of 'Disk Errors'
    and ''ViaRaid Errors' in my 'Event Viewer' for Drive K.
    And with the errors, came frequent data corruption for
    certain files on Drive K. (could not open, file size was
    usually SMALLER than what I know I put on there)
    I was running it this way, cause if I ever had a failure in
    on of the Sata disks I wouldnt lose ALL 500 gigs of
    storage I had like I would if they were in a raid.

    Because of the constant errors on Drive K, I wanted
    to be able to test it and see if it was perhaps a bad
    drive so I went to the Maxtor site to look for a diagnostic
    program, and found that the program (PowerMax) would
    not work and could not even see a Sata drive using an 'imbedded
    controller' .... and even mentioned specifically the Via
    KT800 chipset. Net result, no way to test the individual
    drive(s).

    I went ahead, deleted the two drives, and this time
    installed them, and formatted as a JBOD raid config, still
    using the VIA controller. (cause I see no way to use the
    promise controller in a JBOD config).

    I copied all the data back onto the raid (the same two Maxtor
    drives) and so far, I havent gotten one error in event viewer
    and havent had a single file be corrupted, being unreadable.

    My question is, what am I to make of all this? Is it possible
    that the same drives, same controller, can function okay
    in one kind of set up (raid) and not in another? (non raid)

    Thanks
     
    Mtimerding, Oct 13, 2005
    #1
    1. Advertisements

  2. Mtimerding

    Paul Guest

    Eons ago, there were problems with disk caches at shutdown.
    A lucky user would save some work, just before shutting down
    the computer. For some reason, there was no practical way to
    flush the cache on the disk controller before powering down,
    and a portion of the user's file would be lost when the power
    went off. I think one workaround, was a patch added some delay
    time to shutdown, in an attempt to let the disk controller's
    write-out policy to drain any outstanding writes still left
    in the cache. In some cases, the safest alternative, was to disable
    caching. (Caching can exist at the driver level, such as
    when using the chipset maker's disk driver, or caching
    exists on the cache chip on the disk controller board. The
    disk controller board is the most likely to cause the problem.)

    The difference might be whether an ATA/ATAPI type
    driver (for want of a precise technical term) or a SCSI
    miniport driver is being used. The difference between these
    drivers, just might be the way that the cache is treated. You
    may find one type of driver disables the cache, and that is
    why it works better.

    You could set up the bad config again, save a file with known
    file name just before shutdown, and see if that file is
    corrupted when you next reboot. That would imply a problem
    with cache flushing at shutdown.

    Paul
     
    Paul, Oct 13, 2005
    #2
    1. Advertisements

  3. Mtimerding

    Mtimerding Guest


    Thanks for the reply, and possible explanation. But
    I am curious, in either set up I used the VIA Raid controller,
    so it wouldnt be an issue of "....the way the cache is treated
    based on the driver used" .... would it?
     
    Mtimerding, Oct 13, 2005
    #3
  4. Mtimerding

    Paul Guest

    The SATA cable doesn't know what mode the rest of the chip
    is running in, so about the only other failure mechanism I
    know if, is a cache/power_off mechanism. To prove or disprove
    a cache based theory, all you need to do is write one file to
    disk, just before shutdown. If the file is found to be truncated
    on the next boot cycle, then the cache policy was the culprit.
    _If_ a Microsoft driver is installed, the cache policies could
    be different than if another driver supplier is involved.

    Like all hypothesis, you have to test it :)

    Paul
     
    Paul, Oct 14, 2005
    #4
  5. Mtimerding

    Mtimerding Guest

    Thanks again for the responses Paul .. as I no longer have
    the backups of the 500GB of data, and it took me a heck of
    a long time to backup/restore to what I got now .... I guess
    I will forgo the test for now and will just keep running this way
    (with a JBOD) for now (as long as it keeps working) ....at
    some point in the future, I am sure
    I will end up trying your test ....if nothing else, at some point
    I plan on changing these two 300gb drives for two 500gb (or plus)
    drives ... that will be a perfect time to try it, especially considering
    that running two 500gb in a raid means I would risk LOSING a terabyte
    of data should one drive fail, and definately would NOT want to be
    running in a JBOD raid just for storage. In the meantime, MAYBE
    MAXTOR will release their Powermax (diagnostic) software that
    WILL work on a KT800 chipset and 'imbedded' controllerers.
    As it stands now, sorta sucks not being able to test the harddrive.
    If not, when I go to purchase the bigger harddrives I will definately
    buy from a manufacturer whose disagnostics program will work
    on this rig. (I knew there had to be a reason the Maxtor drives
    were so inexpensive) :) Or maybe some 'genius' will come
    up with a program that allows you to 'save' the data on the other
    disk should one disk fail. (geeesh, I've even learned that my
    defrag program (diskkeeper) will not defrag a drive that's bigger
    than 500+ gb. I guess that's another reason NOT to use a JBOD
    raid)

    Too bad they don't make an external harddrive big enough to transfer
    all this data to (to hold it temporarily while I try the changes)
    I went out looking for one for just that reason, and apparently there
    are none bigger than 250gb - 300gb.

    At least now I know it COULD have been the controller so I am now
    somewhat less confused as to what could have happened.

    Thanks Again
     
    Mtimerding, Oct 14, 2005
    #5
  6. Mtimerding

    Paul Guest

    Hitachi has a 7K500. I'd give a link but the URL might
    not be valid for you, so look in the desktop section
    here ( http://www.hitachigst.com ).

    It looks to be an IDE drive, with $324 being the low price.

    http://castle.pricewatch.com/s/search.asp?s=7k500

    Slap it in an IDE to USB enclosure, and you're done :)

    You can also search by "disk capacity" here, and I see
    that Seagate also has a 500GB drive available.

    http://www.pricewatch.com/m-26.htm

    Paul
     
    Paul, Oct 14, 2005
    #6
    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.