Help...Hung services, unstoppable services, little log information...

Discussion in 'Windows Small Business Server' started by ThereIsNoTry, Aug 26, 2005.

  1. ThereIsNoTry

    ThereIsNoTry Guest

    Hello All:

    What seemed a minor issue on a server this AM seems to be growing in
    seriousness:

    While checking on routine user stuff I noticed that the SBS2000 box (Dell
    PowerEdge 4600, 2 RAID arrays - Data/OS, 1gb RAM, tons of HD space, fully
    patches as per MS, Exchange, IIS, dual-nic) was reporting in the console
    that the APC Agent and APC Server services were in the "starting" mode. The
    box had been up for weeks (at least since the last round of updates) and I
    couldn't immediately verify when they might have stopped and/or been asked
    to start again.

    Attempts at manually restarting ended in "Error 1053: The service did not
    respond to the start or control request in a timely fashion." This happened
    after repeated tries. The event logs showed no relevant information so,
    while the shop was empty, I opted for a quick reboot. At some point during
    the reboot process (after the desktop closed, that much I know for sure) the
    BSOD came up with a stop error:

    Stop: 0x000000D1 (0x...2, 0x...5, 0x...0, 0xF6431951)
    Address F6431951 Base at F6430000 Datestamp 3cd297b9 = perc2.sys

    After creating the huge dump file it rebooted.

    The system came up normally - at least to the login. When I logged in, the
    process took significantly longer than usual and, low and behold, not only
    were the APC services hung in "starting" mode but three BackUp Exec 8.x
    services were in the same condition.
    Manual restarts on those services gave the same 1053 error.
    I tried stopping the services and they switched to hang in the stopping
    state.
    On a hunch, I picked another service to see if I could stop it....The
    Shavlik HFNETchk services also hung in the stopping mode.

    The only related error logged by the eventviewer was for the "savedump":

    Event Type:Information
    Event Source:Save Dump
    Event Category:None
    Event ID:1001
    Date:8/26/2005
    Time:8:46:13 AM
    User:N/A
    Computer:SBSSVR04
    Description:
    The computer has rebooted from a bugcheck. The bugcheck was: 0x000000d1
    (0x00000002, 0x00000005, 0x00000000, 0xf6431951). Microsoft Windows 2000
    [v15.2195]. A dump was saved in: C:\WINNT\MEMORY.DMP.

    No other errors or issues. The users are oblivious to the problem. They are
    printing, exchanging, accessing files without any issue.

    Googling resulted in this: http://support.microsoft.com/?id=307309 Some of
    the symptoms match what's going on but the key fix is not applicable. It
    recommends changing the startup type of the Logical Disk Manager service to
    "Manual" instead of "Automatic" via the registry. My LDM is set to
    "Disabled" ! This is (I think) because Dell OpenManager is handling the
    RAID array's. But this is the closest thing to a clear resolution path that
    I have.

    Anyone have any suggestions on how best to proceed? The weekend is almost
    here so I figured I'd wait so that my users can get in a full day of work
    before I start mucking around.

    Thanks in advance:

    Christopher*
     
    ThereIsNoTry, Aug 26, 2005
    #1
    1. Advertisements

  2. ThereIsNoTry

    ThereIsNoTry Guest

    Leythos:

    Thanks for the rapid response.
    On your suggestion, I visited APC and found the article regarding updating
    due to the Java version running out. Is this the problem you where referring
    to - answer ID 7202?

    It sounds like a pretty good lead. Seems weird that it would kill other
    services as well. I'll follow up.
    Thanks.

    christopher*
     
    ThereIsNoTry, Aug 26, 2005
    #2
    1. Advertisements

  3. ThereIsNoTry

    ThereIsNoTry Guest

    Leythos:

    Thank you, thank you, thank you!

    Your long distance diagnosis was right on point. That was one of the most
    bizarre issues I've seen in more than a decade of babysitting servers and I
    kept picturing tortuous hours of troubleshooting but your quick, detailed
    reply put me on the right track. It took me longer to kick my users off the
    server then it did to make the fix thanks to your input.

    Next time you're in Philly let me know. I owe you a good lunch!

    christopher*
     
    ThereIsNoTry, Aug 29, 2005
    #3
    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.