CLOSE_WAIT Problem. - Weblogic

This is a discussion on CLOSE_WAIT Problem. - Weblogic ; Hi, The following is among the list of fixes in Weblogic 4.5.1. " CR 37789: Sockets on the server were opened by the web server and then cl= osed. But they still were stuck in the CLOSE_WAIT state on the ...

+ Reply to Thread
Results 1 to 4 of 4

CLOSE_WAIT Problem.

  1. Default CLOSE_WAIT Problem.


    Hi,

    The following is among the list of fixes in Weblogic 4.5.1.

    " CR 37789: Sockets on the server were opened by the web server and then cl=
    osed.
    But they still were stuck in the CLOSE_WAIT state on the server. Eventually=
    all
    these sockets took up all the file descriptors available and the server han=
    ged--
    unable to answer requests. This has been fixed."

    We are using Weblogic 6.1 sp2 with aix 433. But we still have the above p=
    roblem.
    How do we handle this?

    Regs
    Vaidy


  2. Default RE: CLOSE_WAIT Problem.

    We are using WebLogic 6.0 SP2 on Solaris 7 and we have a similar problem. =
    A sudden burst of traffic will cause a large build-up of CLOSE_WAIT connect=
    ions on the WebLogic server, and meanwhile on the Apache web serve we see l=
    ots of the following messages in the error_log file:

    [Mon Mar 18 14:39:57 2002] [error] READ_TIMEOUT [os error=3D0, line 485 of=
    ap_proxy.cpp]:

    Seems like a WebLogic/Apache plug-in problem.

    Note: we've tried using Netscape/iPlanet on the web tier, and that results =
    in a build up of the CLOSE_WAITs on the web tier instead of the app tier. =
    Very strange....

    Any ideas?

    Thanks,

    /Greg


  3. Default Re: CLOSE_WAIT Problem.

    Fixed in Weblogic 6.1 sp 2

    044988 Fixed a problem with the NSAPI plug-ins where the socket was not pro=
    perly closed by the plugin side. The un-closed socket on the webserver side=
    will be left in CLOSE_WAIT state.


    Greg Schuler wrote:

    > We are using WebLogic 6.0 SP2 on Solaris 7 and we have a similar problem.=

    A sudden burst of traffic will cause a large build-up of CLOSE_WAIT conne=
    ctions on the WebLogic server, and meanwhile on the Apache web serve we see=
    lots of the following messages in the error_log file:
    >
    > [Mon Mar 18 14:39:57 2002] [error] READ_TIMEOUT [os error=3D0, line 485 =

    of ap_proxy.cpp]:
    >
    > Seems like a WebLogic/Apache plug-in problem.
    >
    > Note: we've tried using Netscape/iPlanet on the web tier, and that result=

    s in a build up of the CLOSE_WAITs on the web tier instead of the app tier.=
    Very strange....
    >
    > Any ideas?
    >
    > Thanks,
    >
    > /Greg




  4. Default RE: Re: CLOSE_WAIT Problem.

    Is the fix in the WL plugin 6.1 Sp2
    effective if we continue to run WL 6.0 Sp2 on our application servers? Ob=
    viously the plugin is easy to update. Upgrading WebLogic on our app server=
    s from 6.0 to 6.1 is significantly more involved....


+ Reply to Thread