A task is an executable that is internal to Adaptive Server (for example, user connections, daemons such as the housekeeper task, and kernel tasks such as I/O handling). Adaptive Server schedules tasks to threads, which assign CPU time to each task.
Adaptive Server tasks include:
User tasks – represent user connections and are capable of executing user queries. An isql connection is an example of a user task.
Service tasks – not associated with specific users, and do not issue user queries. Service tasks include the housekeeper tasks, checkpoints, replication agent threads, and so on).
System tasks – kernel-level versions of service tasks. System tasks include I/O handlers, the clock handler, cluster membership service, IP link monitor, and so on.
Adaptive Server multiplexes tasks across the threads. Run-to-completion (RTC) thread pools place a task in a thread pool and the next available thread picks up the task, which remains on the thread until it completes, and terminates when it finishes running.
See “Base priority” in Chapter 4, “Distributing Engine Resources,” in the Performance and Tuning Series: Basics for a description of setting the task priority for an individual task.
The monTask monitoring table includes a row for each task running on Adaptive Server. See the Reference Manual: Tables.
