Previous | Notification Service Configuration | Next |
The Notification Singleton exists as a single object within a given instance of the Notification Service providing the core service functionality
When the write-ahead log is enabled, information that is normally written to the underlying database is written to a log file instead. When the log file reaches a specific size (defined by the Write Ahead Log Maximum Size property), the database is updated and the log file is reused. The location of the log file is defined by the Write Ahead Log Directory property.
The write-ahead log may increase performance when persistent events are required, particularly when events are being delivered quickly (when consumers are available and responding quickly).
The write-ahead log is enabled when this property is set TRUE (checked).
The directory used to contain write-ahead log files. This directory must be local to the host running the service. The default location is:
<INSTALL>/domains/<domain>/<node>/NotificationService/data
where <INSTALL> is the OpenFusion installation path. See The Object Hierarchy for details of the domains directory structure.
The maximum number of entries that can be stored in the write-ahead log before flushing (writing to the underlying database) takes place.
This property is used when a database plugin is available to OpenFusion to enhance the event persistence mechanism. Leave this field blank when the plugin is not available.
The General properties are useful for setting the start-up parameters of a Notification Service Singleton object. These properties are all static and mainly read -write. All these properties are optional, but can only be set prior to starting the Notification Service Singleton.
The Naming Service entry for the Singleton.
The IOR URL property specifies the location of an Interoperable Object Reference (IOR) for the Service, using the Universal Resource Locator (URL) format. This information is used when a client attempts to resolve a reference to the Service. Currently only http and file URLs are supported, for example:
file:/usr/users/openfusion/NotificationSingleton.ior
http://www.prismtechnologies.com/openfusion/NotificationSingleton.ior
The IOR File Name option specifies the name and location of the IOR file for the Singleton. If this property is not set, the IOR file name will be:
<INSTALL>/domains/<domain>/<node>/<service>/<singleton>/<singleton>.ior
where <INSTALL> is the OpenFusion installation path. See The Object Hierarchy for details of the domains directory structure.
The name of the Naming Service which will be used to resolve the Singleton object.
The ORB Service resolution name used to resolve calls to the Singleton.
The location of the service log file. Each individual component logger (the scheduler logger, the transaction manager logger, and so on) writes to the same service log file. By default, this is the same log file used at the Service level.
The default location of the service log file is:
<INSTALL>/domains/OpenFusion/localhost/NotificationService/ log/NotificationService.log
The format for entries in the service log file. The default format is:
%{priority} [%{category}] %{time:yyyy-MM-dd' 'HH:mm:ss.SSS} %{message}\n%{throwable}
The same format is used by each component logger. This format overrides the format specified in the Log Pattern property at the Service level.
Each component of the Notification Service (the scheduler, the transaction manager, and so on) has its own individual logger. For convenience, every component logger can be set to the same level using this property. Options are:
The default level is Set Individually.
For fine-grained control over logging, set this property to Set Individually. This allows each individual logger to be configured using the individual properties on this tab (described below).
The logger level for the scheduler. Options are:
The logger level for the role manager. Options are:
The default level is Warning.
The logger level for JTO. Options are:
The logger level for the messenger. Options are:
The default level is Warning.
The logger level for the task manager. Options are:
The logger level for the ORB. Options are:
The default level is Warning.
The logger level for the transaction manager. Options are:
The logger level for the blobstore. Options are:
Logger level for the blobstore
The logger level for the state factory. Options are:
The logger level for the state machine factory. Options are:
The default level is Warning.
The logger level for the thread pool. Options are:
The logger level for the event channel factory (which is the root object of the Notification Service). Options are:
The default level is Warning.
The logger level for the component manager. Options are:
The logger level for the lock set factory. Options are:
The default level is Warning.
The interfaces for setting the instrumentation properties are given below. For information on managing instrumentation, please refer to the System Guide.
Instrumentation is not implemented in the initial version 4 release. These properties appear in the Administration Manager but are all locked (disabled). They are described here for completeness.
This property monitors the total number of structured push events received by the Notification Service during execution of the service. In other words, the count of events sent by push suppliers via proxy push consumers.
This property monitors the current number of structured proxy push consumers in existence on the service.
This property monitors the current number of sequence proxy push consumers in existence on the service.
This property monitors the total number of structured push events delivered by the Notification Service during execution of the service. In other words, the count of events received by push consumers via proxy push suppliers.
This property monitors the current number of consumer admins in existence on the service.
This property monitors the current number of structured proxy pull supplier objects in existence on the service.
This property monitors the current number of sequence proxy pull supplier objects in existence on the service.
This property monitors the total number of events in channels.
This property monitors the total number of events awaiting delivery. This count gives the current load on the Service.
This figure is calculated as follows:
Events in queues + (Events in channel * Number of proxies)
This property monitors the current number of structured proxy push supplier objects in existence on the service.
This property monitors the current number of sequence proxy push supplier objects in existence on the service.
This property monitors the current number of Typed Supplier Admins in existence on the service.
This property monitors the current number of unavailable push consumer objects in existence on the service.
This property monitors the current number of unavailable pull supplier objects in existence on the service.
This property monitors the current number of Supplier Admin objects in existence on the service.
This property monitors the current number of Typed Proxy Pull Consumer objects in existence on the service.
This property monitors the current number of Typed Proxy Push Consumer objects in existence on the service.
This property monitors the current number of Typed Proxy Pull Supplier objects in existence on the service.
This property monitors the current number of Typed Consumer Admins in existence on the service.
This property sets the threshold over which delivery queues will raise alarms. The threshold should be expressed as a percentage of the queue's maximum size. So to raise an alarm when a queue is three-quarters full, enter 75.
When this threshold is reached, a DeliveryQueueAlarm event containing the ID of the relevant queue is sent.
This property monitors the current number of Typed Proxy Push Supplier objects in existence on the service.
This property monitors the current number of Event Channel objects in existence on the service.
This property monitors the current number of Typed Event Channel objects in existence on the service.
This property monitors the current number of Structured Proxy Push Consumer objects in existence on the service.
This property monitors the current number of Sequence Proxy Pull Consumer objects in existence on the service.
This property monitors the current number of Structured Proxy Pull Consumer objects in existence on the service.
The maximum queue size of the event delivery manager. When the maximum queue size is exceeded, events are removed from the queue, oldest first, if the EventReliability QoS is set to BestEffort. In the case of Persistent, the events are stored and re-sent when appropriate.
The fairness policy for the lock set factory. Options are:
Although JVM is shown as an option, it is not implemented in the initial version 4 release. FIFO will be be used, regardless of which option is selected for this property.
The frequency with which the task manager reschedules tasks, expressed in milliseconds. The default is 1000 milliseconds (1 second).
This property specifies the number of tasks that the task manager will attempt to keep running at any point in time. This acts as a "soft" limit on thread resource usage. The default number is 10.
The priority for the task manager's own thread. This must be in the range 1 - 10. The default value is 6.
The minimum pool size for the thread pool. The default is 0 (zero).
The maximum size of the thread pool. The default is 20.
The initial size for the thread pool. The default is 0 (zero).
How long, in milliseconds, an idle thread remains in the pool before being discarded. This controls how long an The default timeout is 1000 milliseconds (1 second).
The maximum time allowed before a transaction times out, in milliseconds. The default timeout is 60000 milliseconds (60 seconds).
Previous | Notification Service Configuration | Next |