Livelink Agent Controller Run, How It Works

Adding this case to explain how the background notification agents execute.

1. A database trigger inserts a row into NotifyEvents when an item is added to Livelink.
2. The Agent Controller  will run on a cofngiured schedule. (By default five minutes as defined in the [notify] section of the opentext.ini file.)
3. Two threads are spawned to allow the agents to execute. The Notify thread executes the agents related to notifications.
4. Agent 8999 (aka ClearNotifyEvents agent) processes the events in NotifyEvents, queues events into the LLEventQueue table then deletes the item from the NotifyEvents table.
5. Agent 9000 will call the NodeEventProcessor agent(s) to process the events in LLEventQueue into rows for the NotifyMessages table. Rows in NotifyMessages are visible from the user's (Personal | Notifications ) report screens. The NodeEventProcessor agent is 9001 when not using split notifciation architecture or 900[1234..n] when using split notificaiton architecture. By default, the NodeEventProcessor will process a maximum of 1000 events per run. This is woefully inadequate in a busy system. At Phillips66/CoP, for example, we use 10,000 per pass *AND* configure split notification architecture.
6. Agent 9999 will send notifications by email when a user notification report has email enabled (either because of a system wide default or the user has enabled email delivery). The messages tied to that report in NotifyMessages are sent as a single email message to the user. If the user subscription generates a notification report larger than the configured maximum size allowed, their notificaiton report will warn them of this and truncate the report sent.