Mac Pro 2008 SLS conflict with Sil3132 eSATA RAID card?

Discussion in 'Mac Pro' started by fhturner, Mar 2, 2015.

  1. fhturner macrumors regular

    Joined:
    Nov 7, 2007
    #1
    Hey Everybody—

    I've just set a client up w/ a 2008 Mac Pro running Mac OS X 10.6 Snow Leopard Server. This server also runs Retrospect, backing up to eSATA hard drives. Here are all of the specs:

    Mac Pro 2008
    2.8GHz 8-core
    8GB RAM
    2x1TB internal RAID1
    Mac OS X 10.6.8 Server
    Retrospect v11.5.3

    This eSATA PCIe card: PCIe x1 2-port eSATA RAID w/ Silicon Image 3132 chip
    Driver version 1.7.5.0 (from Silicon Image site)
    1.5TB Fantom Green Drives w/ USB/FW/eSATA​

    When ejecting and disconnecting one backup drive to switch to the other disk, the machine hangs, then eventually restarts itself, as per the Energy Saver settings. Obviously, for a server machine, this is a Particularly Bad Thing™.

    I wondered if the kernel mode might have something to do with it, so yesterday I changed from 32-bit mode to 64-bit mode; no difference. This system should have already been in 64-bit mode by default, according to Apple docs, but it was not.

    I have this same card working w/o issue in a similar setup, but w/ the following differences:

    Mac Pro 2006, 2.66GHz Quad w/ 3GB RAM [32-bit kernel]
    LaCie quad-interface backup drives
    Should I just ditch this eSATA card and try another? I hate to go to that trouble and then find out it's something to do w/ the drive enclosures or system/kernel/driver. Does anyone have any anecdotal info about these components working together? Or any ideas why the card works fine in the other setup but not this one?

    Thanks!
    Fred
     
  2. NOTNlCE macrumors 6502a

    NOTNlCE

    Joined:
    Oct 11, 2013
    Location:
    DMV Area
    #2
    I'm interested in your findings here. I'm looking for an eSATA card myself, and $11 is quite the price point. Have you tried swapping the functioning card into the 2008 machine, on the off chance that you happened to receive a bad card? Perhaps another $11 for a new card is all you need to avoid the Particularly Bad Thing™. :D
     
  3. fhturner thread starter macrumors regular

    Joined:
    Nov 7, 2007
    #3
    Good thoughts, NOTNICE, but both Mac Pros and their eSATA cards are "production" servers, so trying the other card isn't really feasible. And while it's certainly true that I *could* have gotten a bad card, I'm not crazy about the prospect of: ordering another, waiting for it to arrive, taking it to the client site to install, waiting for the client to live-test it and then tell me we're in the exact same boat. I was kinda hoping my problem description might ring a bell for someone here, or that there might be some conventional wisdom that would shed some light on what's going on w/ it. If all else fails, I've got yet another client's 2008 Mac Pro 8-core running SLS that has the NewerTech USB3 + eSATA card installed, and that works fine; so I guess I could order that, despite it being a bit of overkill vs. this $11 card that's working elsewhere.

    Thx,
    Fred
     
  4. TzunamiOSX macrumors regular

    Joined:
    Oct 4, 2009
    Location:
    Germany
    #4
    I would say: looks like the card does not support hot-plug.
     
  5. fhturner thread starter macrumors regular

    Joined:
    Nov 7, 2007
    #5
    Wouldn't that be a function of the Sil3132 chipset on the card? And if the other identical card in the 2006 Mac Pro hot-plugs/swaps w/o issue, shouldn't this one as well?
     
  6. fhturner thread starter macrumors regular

    Joined:
    Nov 7, 2007
    #6
    It never occurred to me before, but when I was looking at info for a different eSATA card, I was reminded of a chipset incompatibility w/ some drive enclosures and Mac OS X via eSATA. Those drives that use the Oxford OXU931DS chipset cause kernel panics or lack of recognition of the drive by the OS. See Technical Note #1 under Technical Specifications tab on this product page:

    http://www.sonnettech.com/product/tempo_sata_e4p.html

    That sounds exactly like what I'm seeing, although I think the Fantom drives we used there are triple- or quad-interface (including FireWire). Perhaps the same chip is used in parallel w/ a FireWire chip. I intend to open the case and have a look when I'm there next, hopefully this week.
     

Share This Page