Slow performance when dumping core

From: Fredrik Robertsson (frero_at_swip.net)
Date: 05/27/04

  • Next message: Anamika Dasgupta: "to caputer the output of explorer"
    To: <sunmanagers@sunmanagers.org>
    Date: Thu, 27 May 2004 10:48:48 +0200
    
    

    Hi,

    I have a problem with two machines: when the application crashes and wants
    to dump core the disk write speed goes down to about 1MB/s as opposed to the
    normal 100-200MB/s of these devices. So the slow speed only shows its ugly
    head when writing the corefiles to disk. I have no problems whatsoever with
    "normal" I/O-operations. When testing disk performance I have no problem
    exceeding the 200MB/s theoretical maximum, and most of the time I get
    somewhere between 100-200MB/s (depending on file size and I/O-request size,
    so that is buffer/cache related).

    The machines are Sun Fire 240 and 480. Both of them have 2 CPU's and 4GB of
    memory. They both have JNI FCE-6460-N FC HBA's, and they are both connected
    to a Hitachi 9980V through two Brocade 12000. They are running Solaris 8
    (one with kernel 117350-02 and the other with 108528-29) with Veritas
    Foundation Suite 3.5 MP1 and version 5.3.0.1 of the JNI driver.

    The application is developed in-house and has a memory leak somewhere. So
    when it reaches 4GB of allocated memory it crashes and tries to dump core.
    The developers says it's tricky to find the leak unless they get a corefile,
    and with 1MB/s that's over an hour of downtime which we cannot allow for.

    Anyone out there had the same problem as poor me?

    Med vdnlig hdlsning/Regards,

    Fredrik Robertsson
    Systems Architect
    Network Engineering
    Tele2 Sverige AB
    +46-(0)8-5626 4000
    _______________________________________________
    sunmanagers mailing list
    sunmanagers@sunmanagers.org
    http://www.sunmanagers.org/mailman/listinfo/sunmanagers


  • Next message: Anamika Dasgupta: "to caputer the output of explorer"

    Relevant Pages

    • Proposed Assembler Commands
      ... ACM Automatically Clear Memory ... BKCRDR Backspace Card Reader ... BKSPD Backspace Disk ... EIAO Execute In Any Order ...
      (sci.electronics.design)
    • Spontaneous reboots
      ... Postfix, Cyrus-Imap and Apache are all configured with TLS ... The crashes _always_ causes disk errors that cannot be recovered ... I have now deinstalled all ports. ...
      (freebsd-questions)
    • Re: Att. Alex Nichol -VM cont.
      ... I find documentation that says a 4kb page in memory is written to the hard ... manner that a 4kb memory page is equal to a 4kb cluster written on the disk". ... So where is the basis for saying since 4kb paging in memory that 4kb clusters ...
      (microsoft.public.windowsxp.general)
    • Re: teaching a child - console or GUI
      ... I use a RDBMS to reuse existing code and optimizations. ... >>And loading them in a system under load (with constant disk io) is worse ... quality stuff), and quality power supply, quality memory etc. ... > lists of pointers on disk, extract files of frequent search fields in ...
      (comp.lang.pascal.delphi.misc)
    • Re: First-timer buying AS/400 and have questions
      ... > (writes memory to disk for the next boot). ... > the 170 has a battery, which you will probably need to replace it ... Ahh, makes a little more sense now how Silverlake relates to Rochester, ...
      (comp.sys.ibm.as400.misc)