IBM Support

PH48686: OVERRIDE STUCK CONNECTION SUPPORT SECONDS WITH MILLISECOND UNITS

Subscribe to this APAR

By subscribing, you receive periodic emails alerting you to the status of the APAR, along with a link to the fix after it becomes available. You can track this item individually or track all items by product.

Notify me when this APAR changes.

Notify me when an APAR for this component changes.

 

APAR status

  • Closed as new function.

Error description

  • The administrator console for stuck connection support only
    allows for units in seconds.  Customers may be finding that
    the units are not small enough for detecting stuck connections
    fast enough on high performance systems and require the use of
    millisecond instead.
    
    A new custom connection pool property stuckUnitsMilliseconds
    boolean value will be provided to use milliseconds when set to
    true.  The default of seconds will remain if
    stuckUnitsMilliseconds in not used or stuckUnitsMilliseconds =
    false
    

Local fix

Problem summary

  • ****************************************************************
    * USERS AFFECTED:  All users of IBM WebSphere Application      *
    *                  Server                                      *
    *                  using stuck connection detection.           *
    ****************************************************************
    * PROBLEM DESCRIPTION: Current stuck connection detection uses *
    *                      application connections to determine    *
    *                      when                                    *
    *                      connections are stuck.                  *
    ****************************************************************
    * RECOMMENDATION:                                              *
    ****************************************************************
    The existing stuck connection support relies on existing
    connection in the connection pool for determining stuck
    connections.  An  enhancement to the existing stuck connection
    support has been added  to the service stream.
    The enhancement will use a connection outside of the
    application connections being used to determine stuck
    connections.
    This uses the amount of time it takes a connection to be
    created, closed, cleaned up and destroyed without using a
    query.  This enhancement is targeted for customers who may
    have tried using the existing  stuck connection support but
    could not get a reliable stuck  timing due to the application
    usage of connections.
    Both configuration and mbeans were added for this function.
    Custom connection pool properties added stuckQueryTime,
    stuckQueryInterval, stuckQueryAverageInterval, and
    stuckQueryDetectionInterval with the same names and units being
    used for the attribute mbeans added.
    attribute name="stuckQueryTime"
    description="The amount of time connection can be stuck before
    switching to stuck mode true"
    units=milliseconds
    attribute name="stuckQueryInterval"
    description="Frequency of stuck query time thread when not in
    stuck
    mode true"
    units=milliseconds
    attribute name="stuckQueryAverageInterval"
    description="The number of past stuck duration times to use
    for an
    average stuck duration time"
    units=number of values for to average.
    attribute name="stuckQueryDetectionInterval"
    description="Frequency of stuck query time thread when in stuck
    mode true"
    units=milliseconds
    For both the existing and the enhancement to stuck connection
    support, when time in use exceeds the stuck time, the stuck
    mode
    will switch to true.   When stuck mode is true, the connection
    pool
    will throw an exception to any new connection request until
    stuck
    mode is returned to false.   For the stuck mode to return to
    false,
    the stuck time being used must drop below the stuck time.
    For the enhanced behavior, stuckQueryAverageInterval and
    stuckQueryDetectionInterval have been added to help with
    timing of
    the create, close, cleanup and destroy that vary large amounts
    or
    move in and out of stuck mode often.
    The stuckQueryAverageInterval will average the times over x
    amount
    of them.  Example if 10 is used, the last 10 values will be
    averaged and used to determine stuck mode.
    The stuckQueryDetectionInterval is used instead of the
    stuckQueryInterval during stuck mode true.   This can slow
    down the return to stuck mode false if this is required for
    better return to normal.  Example, when stuck mode is false,
    the stuckQueryInterval  wil be used and when stuck mode is
    true, stuckQueryDetectionInterval will be used for the
    interval.
    Uses for this function does vary from customer to customer.
    Most customers who use this function want to be able to have
    application automatically switch to use some other resources
    that are not stuck.   Any application getting the stuck
    connection exception can switch to other resources.  Other
    customer may just want some  relief to the back-end resource
    that has slowed down, with the best example of have 10 servers
    running the same applications to the same resources.  Since
    this stuck connection detection is active on
    each server, one server may enter stuck more, while other
    continue to use resources.   This is just a few examples for
    this advanced connection pool function.
    Some existing customers may have already implemented function
    that already provides similar behavior using an application
    that monitors connection behavior.  Example, when the customer
    application determines a resource is slow or stuck, it uses the
    connection pool mbean pause to prevent new connection requests
    from being completed and then resumes the connection pool when
    itâ??s determined that the resource is back to normal.   The new
    enhanced stuck connection detector may allow some customers
    with existing  simple stuck detection function applications to
    maintain less code  and switch to use this built in enhanced
    stuck connection detection function.
    Minimum enablement of the enhanced stuck detection is
    stuckQueryTime and stuckQueryInterval must be set.
    

Problem conclusion

  • The stuck connection enhancement has been added.  Problem
    summary contains both problem summary and conclusion.
    
    
    The fix for this APAR is targeted for inclusion in fix pack
    9.0.5.15. For more information, see 'Recommended Updates for
    WebSphere Application Server':
    https://www.ibm.com/support/pages/node/715553
    

Temporary fix

Comments

APAR Information

  • APAR number

    PH48686

  • Reported component name

    WEBSPHERE APP S

  • Reported component ID

    5724J0800

  • Reported release

    900

  • Status

    CLOSED PER

  • PE

    NoPE

  • HIPER

    NoHIPER

  • Special Attention

    NoSpecatt / Xsystem

  • Submitted date

    2022-08-15

  • Closed date

    2023-03-17

  • Last modified date

    2023-03-17

  • APAR is sysrouted FROM one or more of the following:

  • APAR is sysrouted TO one or more of the following:

Fix information

  • Fixed component name

    WEBSPHERE APP S

  • Fixed component ID

    5724J0800

Applicable component levels

[{"Business Unit":{"code":"BU059","label":"IBM Software w\/o TPS"},"Product":{"code":"SSEQTP","label":"WebSphere Application Server"},"Platform":[{"code":"PF025","label":"Platform Independent"}],"Version":"9.0","Line of Business":{"code":"LOB45","label":"Automation"}}]

Document Information

Modified date:
18 March 2023