RE: cfdcollect behaving badly

From: K.C. Smith (kace@netrail.net)
Date: Tue Nov 28 2000 - 12:57:49 PST

  • Next message: Andrew Pasetti: "core dump while executing patched cflowd"

    It seems that the problem was one of overloading. When I increased the
    minPollInterval and reduced the number of different data types kept the
    problem stopped.

    K.C.

    -----Original Message-----

    Has anyone seen anything like this before?

    Nov 27 08:56:56 hostname cfdcollect[10787]: [I] cfdcollect (version
    cflowd-2-1-b1) started with 2 cflowd instances.
    Nov 27 08:56:57 hostname cfdcollect[10787]: [I] connected to localhost:2056
    Nov 27 09:06:57 hostname cfdcollect[10787]: [I] awakened by alarm.
    Nov 27 09:06:57 hostname cfdcollect[10787]: [I] localhost has data for
    134742016 routers.
    Nov 27 09:07:13 hostname cflowd[10788]: [I] sent data to 127.0.0.1:1768
    Nov 27 09:07:13 hostname cfdcollect[10787]: [E]
    ArtsPrimitive.FdRead(4,0xbffeb424,81920) failed: Interrupted system call
    {CflowdPortMatrix.cc:135}
    Nov 27 09:07:13 hostname cfdcollect[10787]: [E]
    ciscoInterface.PortMatrix().read(4) failed: Interrupted system call
    {CflowdCisco.cc:459}
    Nov 27 09:07:13 hostname cfdcollect[10787]: [E] cflowdCisco.read(4) failed
    {CflowdServer.cc:711}

    cflowdmux and cflowd always seemed to have started correctly first. In
    particular, I've seen the "cfdcollect[pid]: [I] localhost has data for
    134742016 routers." message several times, always followed by various
    errors. Any ideas would be greatly appreciated.

    K.C. Smith

    --
    cflowd mailing list
    cflowd@caida.org
    

    -- cflowd mailing list cflowd@caida.org



    This archive was generated by hypermail 2b29 : Tue Nov 28 2000 - 13:20:07 PST