On Fri, 2009-02-20 at 09:07 -0500, Carolyn Beeton wrote:
> > Sending the profile indeed results in the following log entry..
> >
> > 2009-02-19T12:49:54.818165Z:3402:SUPERVISOR:WARNING:sun.qantom
> > .int:HttpServer-3:B7C10B90:Supervisor:'ConfigAgent': Received
> > u nexpected event evConfigurationChanged while in state 'Stopping'
>
> This is just the supervisor state machine warning that the event
> evConfigurationChanged is not explicitly handled in state Stopping.
> Since this occurs during normal operation and does not indicate a
> problem, I will lower the log level.
While you're at it, rephrase the message. Configuration change events
should never be 'unexpected' - they are perfectly normal and allowed to
happen at any time in any state.
I recently updated the descriptions of the log levels in
sipXportLib/include/os/OsSysLog.h:
PRI_DEBUG, /**< Developer message needed only when debugging code.
* May include recording of such things as entry and exit
from methods
* or internal branch tracking.
* Should never be required by end users.
*/
PRI_INFO, /**< Informational message used to trace system inputs and
actions.
* Significant actions in the execution of some activity;
for example, the
* receipt of a message or an important decision in its
disposition.
* This level should be sufficient for an administrator or
support person
* to determine what occured in the system when debugging a
configuration
* or interoperability problem in the field.
*/
PRI_NOTICE, /**< Normal but significant events.
* Events that are expected but provide important context,
such as service
* restarts and reloading configuration files.
* This is the default logging level, and generally logging
should
* always include at least these messages.
*/
PRI_WARNING, /**< Conditions that imply that some failure is possible but
not certain.
* Generally, external inputs that are not as expected and
possibly
* invalid. Especially useful in low level routines that
are going to
* return an error that may be recoverable by the caller.
*/
PRI_ERR, /**< An unexpected condition likely to cause an end-user
visibile failure.
* This level should be used whenever an error response is
being sent
* outside the system to provide a record of the internal
data that
* are important to understanding it.
*/
PRI_CRIT, /**< Endangers service operation beyond the current operation.
* MUST be logged prior to any 'assert', or when exiting
for any abnormal
* reason.
*/
PRI_ALERT, /**< Fault to be communicated to operations.
* Should be replaced by usage of the new Alarm subsystem.
*/
PRI_EMERG, /**< System is unusable.
*/
... per the above, I'd recommend making that INFO or DEBUG (on the
assumption that a more specific event describing what the configuration
change was is logged elsewhere at NOTICE, which I believe is true).
_______________________________________________
sipx-dev mailing list
[email protected]
List Archive: http://list.sipfoundry.org/archive/sipx-dev
Unsubscribe: http://list.sipfoundry.org/mailman/listinfo/sipx-dev
