8009 & GPF problems
  • Ron
    8009 & GPF problems

    by Ron » Tue Oct 30, 2007 8:26 am

    Firstly hope all goes well with you guys.

    We are stuck with the following 2 issues:

    A) Once every day or two we get a commix error 8006 or 8009. These are reported in caperror.txt which has been emailed to Martine. We have rebuilt the index's and this still happens.

    B) Approximately once a day we get a capwin.exe, fault in module @ 003C 078D general protection fault. This occurs both on Windows 2000 and windows XP machines.

    We have run your network stress test for 2 sessions of 3 days without errors. we have also ensured there are no disk errors/bad sectors etc. We also did the command line database repair as requested.

    Our usage is very low with the system often idle or with one user out of four.

    Please make some suggestions for a cure.
  • COBS Tech Support
    Posts:683
    Joined:Fri Sep 09, 2005 8:23 am

    by COBS Tech Support » Tue Oct 30, 2007 8:30 am

    Error 8009 is a corrupt index. And 8006 effectively means the same thing.

    If you also see 1**** errors, typically 1010 in the log (CAPERR.TXT), this means the system is breaking the network connection somewhere (read error on network or hard drive). At face value given the limited information available to help you, it's plausible that the GPF's are causing the index corruption. Two things cause these types of problems: software freezing up, when writing to the database, or hardware failure (intermittently faulty network adaptor, switch, router, etc.) For example, a faulty switch might cause such a problem.

    The software freeze-up might be caused by some kind of issue in the OS itself, e.g., corrupted registry settings, printer driver leaking memory, and so on.

    Is there any particular pattern to the problem? I.e., the GPF's seem to start after the use has printed a few hundred statements? (Printer driver issue?)

    Or does it happen at any time of the day completely randomly? (Problems with memory chips, etc.?)

    Once an index is corrupted, indexes need to be rebuilt, otherwise the problems will continue when there is nothing really wrong with the system.

    Run a disk check first on the server. Then execute:

    C:\CAPITAL\INSTWIN.EXE /A

    to rebuild *all* indexes. A common mistake is for users to rebuild the indexes in Company A but they are actually working in Company B.

    What version of Windows are they running?

    There is a utility called NETSAFE.EXE found in the \CAPITAL folder. Run this on the server and on every PC that connects to the server. This will turn off all the speed optimizations in Windows that either corrupt peer-to-peer database sharing, or that are generally unsafe. You do not need to run NETSAFE if you are operating the Corporate Edition of CAPITAL Office.

    The cause of problems like these can sometimes be very challenging to isolate. Which OS are they running? How old is it? Freshly installed? More than 3 years old? If a system has had a lot of software installed and uninstalled on it, that can sometimes causes stability issues later on.

Who is online

Users browsing this forum: No registered users and 26 guests