IO Completion Listener

Collapse
This topic is closed.
X
X
 
  • Time
  • Show
Clear All
new posts
  • abolduc@gmail.com

    IO Completion Listener

    I have had an underwhelming amount of success hunting down the source
    of this error and am hoping that someone here may have some insight.

    Error logged in the event log:
    Source: MSSQLSERVER
    Category: (2)
    EVENT ID: 17887
    Description: IO Completion Listener (0x754) Worker 0x00FEC0E8 appears
    to be non-yielding on Node 1. Approx CPU Used: kernel 0 ms, user 0 ms,
    Interval: 15187.

    Immediately followed by 100% CPU utilization. The server remains ping-
    able and I can telnet to the SQL port, but it is otherwise effectively
    unresponsive (queries timeout, RDP times out, etc.). This is also
    logged in the SQL error log:

    2008-03-19 15:52:37.80 Server Using 'dbghelp.dll' version '4.0.5'
    2008-03-19 15:52:46.80 Server **Dump thread - spid = 0, PSS =
    0x00000000, EC = 0x00000000
    2008-03-19 15:52:46.80 Server ***Stack Dump being sent to D:
    \Microsoft SQL Server\MSSQL.3\ MSSQL\LOG\SQLDu mp0004.txt
    2008-03-19 15:52:46.80 Server *
    *************** *************** *************** *************** *************** ****
    2008-03-19 15:52:46.80 Server *
    2008-03-19 15:52:46.80 Server * BEGIN STACK DUMP:
    2008-03-19 15:52:46.80 Server * 03/19/08 15:52:46 spid 0
    2008-03-19 15:52:46.80 Server *
    2008-03-19 15:52:46.80 Server * Non-yielding IOCP Listener
    2008-03-19 15:52:46.80 Server *
    2008-03-19 15:52:46.80 Server *
    *************** *************** *************** *************** *************** ****
    2008-03-19 15:52:46.80 Server *
    -------------------------------------------------------------------------------
    2008-03-19 15:52:46.80 Server * Short Stack Dump
    2008-03-19 15:58:13.09 Server Stack Signature for the dump is
    0x00000006
    2008-03-19 16:03:40.94 Server IO Completion Listener (0x730)
    Worker 0x00A060E8 appears to be non-yielding on Node 0. Approx CPU
    Used: kernel 0 ms, user 0 ms, Interval: 15187.
    2008-03-19 16:06:10.47 Server Timeout waiting for external dump
    process 6016.

    2008-03-19 16:06:10.47 Server IO Completion Listener (0x754)
    Worker 0x00FEC0E8 appears to be non-yielding on Node 1. Approx CPU
    Used: kernel 0 ms, user 0 ms, Interval: 15187.
    2008-03-19 16:17:34.51 Server IO Completion Listener (0x754)
    Worker 0x00FEC0E8 appears to be non-yielding on Node 1. Approx CPU
    Used: kernel 0 ms, user 0 ms, Interval: 15187.


    The server is Windows 2003 SP2 (32bit), dual dual-core AMD opteron
    processors, 8GB RAM, RAID 10. /3GB and /PAE are enabled in boot.ini.
    "Lock pages in memory" has been granted to the account SQL runs as.

    SQL Server is 2005 Standard Edition SP1 (32bit). "Use AWE to allocate
    memory" is enabled. Min and max memory are set to 5GB.

    If there is any other info you think may be helpful in
    troubleshooting , please let me know. Any insight you've had into this
    or similar problem would be appreciated. Thanks in advance.

  • Tom van Stiphout

    #2
    Re: IO Completion Listener

    On Mon, 24 Mar 2008 09:44:00 -0700 (PDT), abolduc@gmail.c om wrote:

    SQL Server is at SP2. Why not get current first.


    -Tom.

    >I have had an underwhelming amount of success hunting down the source
    >of this error and am hoping that someone here may have some insight.
    >
    >Error logged in the event log:
    >Source: MSSQLSERVER
    >Category: (2)
    >EVENT ID: 17887
    >Description: IO Completion Listener (0x754) Worker 0x00FEC0E8 appears
    >to be non-yielding on Node 1. Approx CPU Used: kernel 0 ms, user 0 ms,
    >Interval: 15187.
    >
    >Immediately followed by 100% CPU utilization. The server remains ping-
    >able and I can telnet to the SQL port, but it is otherwise effectively
    >unresponsive (queries timeout, RDP times out, etc.). This is also
    >logged in the SQL error log:
    >
    >2008-03-19 15:52:37.80 Server Using 'dbghelp.dll' version '4.0.5'
    >2008-03-19 15:52:46.80 Server **Dump thread - spid = 0, PSS =
    >0x00000000, EC = 0x00000000
    >2008-03-19 15:52:46.80 Server ***Stack Dump being sent to D:
    >\Microsoft SQL Server\MSSQL.3\ MSSQL\LOG\SQLDu mp0004.txt
    >2008-03-19 15:52:46.80 Server *
    >************** *************** *************** *************** *************** *****
    >2008-03-19 15:52:46.80 Server *
    >2008-03-19 15:52:46.80 Server * BEGIN STACK DUMP:
    >2008-03-19 15:52:46.80 Server * 03/19/08 15:52:46 spid 0
    >2008-03-19 15:52:46.80 Server *
    >2008-03-19 15:52:46.80 Server * Non-yielding IOCP Listener
    >2008-03-19 15:52:46.80 Server *
    >2008-03-19 15:52:46.80 Server *
    >************** *************** *************** *************** *************** *****
    >2008-03-19 15:52:46.80 Server *
    >-------------------------------------------------------------------------------
    >2008-03-19 15:52:46.80 Server * Short Stack Dump
    >2008-03-19 15:58:13.09 Server Stack Signature for the dump is
    >0x00000006
    >2008-03-19 16:03:40.94 Server IO Completion Listener (0x730)
    >Worker 0x00A060E8 appears to be non-yielding on Node 0. Approx CPU
    >Used: kernel 0 ms, user 0 ms, Interval: 15187.
    >2008-03-19 16:06:10.47 Server Timeout waiting for external dump
    >process 6016.
    >
    >2008-03-19 16:06:10.47 Server IO Completion Listener (0x754)
    >Worker 0x00FEC0E8 appears to be non-yielding on Node 1. Approx CPU
    >Used: kernel 0 ms, user 0 ms, Interval: 15187.
    >2008-03-19 16:17:34.51 Server IO Completion Listener (0x754)
    >Worker 0x00FEC0E8 appears to be non-yielding on Node 1. Approx CPU
    >Used: kernel 0 ms, user 0 ms, Interval: 15187.
    >
    >
    >The server is Windows 2003 SP2 (32bit), dual dual-core AMD opteron
    >processors, 8GB RAM, RAID 10. /3GB and /PAE are enabled in boot.ini.
    >"Lock pages in memory" has been granted to the account SQL runs as.
    >
    >SQL Server is 2005 Standard Edition SP1 (32bit). "Use AWE to allocate
    >memory" is enabled. Min and max memory are set to 5GB.
    >
    >If there is any other info you think may be helpful in
    >troubleshootin g, please let me know. Any insight you've had into this
    >or similar problem would be appreciated. Thanks in advance.

    Comment

    Working...