Getting Started with OpenVMS: A Guide for New Users (HP Technologies)

OpenVMS assigns various states to processes to indicate their eligibility for scheduling. A process is in exactly one state at a given time. Most of the possible process states are as follows:

• CUR (current) This process is currently executing on a CPU. If the system has more than one CPU, the SHOW SYSTEM display will indicate which CPU is executing the process, as in "CUR 06." Your own process will always be shown as current when using SHOW SYSTEM.

• COM (computable) This process is ready to execute and will be placed into execution when a CPU becomes available, subject to priority-based selection. It is waiting for no system resources other than CPU availability.

  When a process in a wait state becomes eligible for processing (the event for which it waited has occurred), it is placed into COM state, where it awaits selection for execution.

• LEF (local event flag wait) The process is waiting for some event that must occur before the process can proceed. Typically, the process is waiting for the completion of a disk I/O operation or keyboard input. A "local event flag" is an indicator within a process that will be triggered when the associated event occurs. A process may wait for a single event or wait for any one of a set of events, awaiting whichever event occurs first.

• CEF (common event flag wait) This is like the LEF state, except that the flag(s) for which the process waits are shared by more than one process. Common event flags are most often used for interprocess synchronization or communication.

• HIB (hibernating) The process has voluntarily placed itself into a wait state from which it will be awakened at a prescheduled time or at the explicit request of another process. Processes that perform some task at regular intervals or that need to execute only when certain system events occur often place themselves into this state while waiting.

• PFW (page fault wait) The process has been stalled by the memory management subsystem. It has made a reference to some part of its address space that is not currently in physical memory (caused a page fault). When the memory management code has made the memory page ready, the process will be resumed at the operation that caused the page fault.

• MWAIT (miscellaneous wait) The process is waiting for one of a number of conditions. Various wait states are grouped together and identified by this category.

• RWxxx (resource wait) Wait states beginning with RW indicate a resource wait, and the subsequent characters describe the particular resource for which the process waits. For example, RWMPB means modified page writer busy, a type of memory management wait, and RWSCS means system communication services, a wait for the Distributed Lock Manager.

• SUSP (suspended) The process has been suspended, most likely by the SET PROCESS/SUSPEND command. It will remain suspended until explicitly resumed by another process.

• COLPG (collided page wait) Another memory management wait. The process has referenced a shared memory page for which another process was already in PFW state.

• Outswapped OpenVMS supports swapping, which is the removal of an entire process from memory. When memory is tight, idle, or nearly idle, processes will first be "trimmed" (have their physical memory usage reduced). Then, if sufficient memory cannot be reclaimed, a trimmed process will be moved entirely out to disk (outswapped).

  A process that is outswapped will be in a state corresponding to the one it occupied at the time it was outswapped. It may be in HIBO (Hibernating, Outswapped) or LEFO (Local Event Flag Wait, Outswapped), for example. When the process becomes eligible for execution, it will be placed into COMO (Computable, Outswapped), and will subsequently be brought into memory (Inswapped) and placed in COM state.