Remoting memory leak?

Collapse
This topic is closed.
X
X
 
  • Time
  • Show
Clear All
new posts
  • Yair

    Remoting memory leak?

    Hi,
    After executing a server-client remoting scheme I wrote,
    I've noticed a constant memory growth (both at the server,
    and at the client). Searching through various newsgroups,
    I've stumbled across various posts, regarding this
    behaviour.
    My question is: Is there currently any known bug,
    regarding Dotnet 1.1, which can lead to such behaviour? We
    have spent a significant amount of time on the subject,
    and ruled out the possibility of having
    something "holding" references to our own objects.
    Thank you very much for your time.
  • Dmitry Baibakov

    #2
    Re: Remoting memory leak?

    Hi,
    what kind of activation do you use?
    If the remoting object is client-activated:
    - Do you override InitializeLifet imeService?
    - Does your client implement ISponsor interface?

    Dmitry.

    Yair wrote:
    [color=blue]
    > Hi,
    > After executing a server-client remoting scheme I wrote,
    > I've noticed a constant memory growth (both at the server,
    > and at the client). Searching through various newsgroups,
    > I've stumbled across various posts, regarding this
    > behaviour.
    > My question is: Is there currently any known bug,
    > regarding Dotnet 1.1, which can lead to such behaviour? We
    > have spent a significant amount of time on the subject,
    > and ruled out the possibility of having
    > something "holding" references to our own objects.
    > Thank you very much for your time.[/color]

    Comment

    • Guest's Avatar

      #3
      Re: Remoting memory leak?

      Hi,
      Thank you for your reply.

      Our remoting object is server-activated (singleton). So, a
      single remoting object should handle all of our remoting
      calls.
      Also, we override 'InitializeLife timeService' and force a
      null return - so the lifetime of the remoting object
      should be infinite.
      Checking the client process, we noticed that numerous
      threads are being opened by the framework, and running
      NetStat at the background, showed that a lot of TCP
      sockets are being opened.
      Further testing showed that the number of threads being
      opened, increases with direct relation to the frequency of
      the remote calls being made by client (In any case, we
      noticed that the number of threads exceeds 25, which to
      our understanding is supposed to be the maximal amount of
      threads in the thread pool).
      [color=blue]
      >-----Original Message-----
      >Hi,
      >what kind of activation do you use?
      >If the remoting object is client-activated:
      >- Do you override InitializeLifet imeService?
      >- Does your client implement ISponsor interface?
      >
      >Dmitry.
      >
      >Yair wrote:
      >[color=green]
      >> Hi,
      >> After executing a server-client remoting scheme I[/color][/color]
      wrote,[color=blue][color=green]
      >> I've noticed a constant memory growth (both at the[/color][/color]
      server,[color=blue][color=green]
      >> and at the client). Searching through various[/color][/color]
      newsgroups,[color=blue][color=green]
      >> I've stumbled across various posts, regarding this
      >> behaviour.
      >> My question is: Is there currently any known bug,
      >> regarding Dotnet 1.1, which can lead to such behaviour?[/color][/color]
      We[color=blue][color=green]
      >> have spent a significant amount of time on the subject,
      >> and ruled out the possibility of having
      >> something "holding" references to our own objects.
      >> Thank you very much for your time.[/color]
      >
      >.
      >[/color]

      Comment

      • Guest's Avatar

        #4
        Re: Remoting memory leak?

        Hi,
        My apologies. It seems that we had a bug in our SW. The SW
        programmer (Let's call him 'Y' for this matter), kept
        registrating events within the thread pool.
        Thank you for your attention.
        [color=blue]
        >-----Original Message-----
        >Hi,
        >Thank you for your reply.
        >
        >Our remoting object is server-activated (singleton). So,[/color]
        a[color=blue]
        >single remoting object should handle all of our remoting
        >calls.
        >Also, we override 'InitializeLife timeService' and force a
        >null return - so the lifetime of the remoting object
        >should be infinite.
        >Checking the client process, we noticed that numerous
        >threads are being opened by the framework, and running
        >NetStat at the background, showed that a lot of TCP
        >sockets are being opened.
        >Further testing showed that the number of threads being
        >opened, increases with direct relation to the frequency[/color]
        of[color=blue]
        >the remote calls being made by client (In any case, we
        >noticed that the number of threads exceeds 25, which to
        >our understanding is supposed to be the maximal amount of
        >threads in the thread pool).
        >[color=green]
        >>-----Original Message-----
        >>Hi,
        >>what kind of activation do you use?
        >>If the remoting object is client-activated:
        >>- Do you override InitializeLifet imeService?
        >>- Does your client implement ISponsor interface?
        >>
        >>Dmitry.
        >>
        >>Yair wrote:
        >>[color=darkred]
        >>> Hi,
        >>> After executing a server-client remoting scheme I[/color][/color]
        >wrote,[color=green][color=darkred]
        >>> I've noticed a constant memory growth (both at the[/color][/color]
        >server,[color=green][color=darkred]
        >>> and at the client). Searching through various[/color][/color]
        >newsgroups,[color=green][color=darkred]
        >>> I've stumbled across various posts, regarding this
        >>> behaviour.
        >>> My question is: Is there currently any known bug,
        >>> regarding Dotnet 1.1, which can lead to such[/color][/color][/color]
        behaviour?[color=blue]
        >We[color=green][color=darkred]
        >>> have spent a significant amount of time on the[/color][/color][/color]
        subject,[color=blue][color=green][color=darkred]
        >>> and ruled out the possibility of having
        >>> something "holding" references to our own objects.
        >>> Thank you very much for your time.[/color]
        >>
        >>.
        >>[/color]
        >.
        >[/color]

        Comment

        Working...