OVNM 5.3 - Operations Agent for NonStop Event Management Guide for UNIX

HP NonStop Event Management System 21
2-2-3 EMS Message Collectors
EMS supports two types of event message collectors.
Primary Collector ($0)
Each system (or node) has only one primary event message collector, named $0. The primary collector, initiated
at system generation, provides a central collection point for all events from all subsystems in a system. $0 uses
the $ZOPR process to perform waited operations. $ZOPR is a continuous process that, like $0, exists on every
node and is initiated at system generation. $ZOPR has no user interface and is used only as described here.
Alternate Collectors
Alternate collectors offer an alternative to the central collection point provided by the primary collector, $0.
Alternate collectors provide functions similar to those supplied by $0, but each alternate collector maintains its
own log files. The separation of events into several log files speeds up event processing because a network
application program does not have to read a single large file containing many events unrelated to that
application.
Some features of alternate collector usage are:
• One or more alternate collectors can be started on a system.
• An alternate collector is typically started by entering a TACL RUN command.
• An alternate collector is a named process (or process pair) with a name you define when you initiate it.
• An alternate collector accepts EMS events sent to it rather than to the primary collector.
• An alternate collector writes tokenized events to a log file. This file is identical in structure to the log files
maintained by the primary collector.
2-2-4 Event Message Collector Functions
The primary and alternate collectors provide:
• Subsystem support
• Pre-log filtration (PLF)
• Log file management
• Distributor support
Subsystem Support
Subsystems and user-written applications send event information to the primary and alternate collectors by means
of the WRITEREAD procedure call. Event messages are based on tokens rather than text.
Pre-Log Filtration (PLF)
Pre-log filtration lets primary and alternate collectors use EMS filters to determine whether specific events should
be logged to the EMS log files. Judicious use of PLF conserves disk space and system resources that might
otherwise be required to write unnecessary or redundant events to the EMS logs.
PLF is implemented by adding one or more filters to a collector:
• Compiled filters and filter tables can fail one or more particular events, thereby preventing the failed events
from being forwarded to the collector’s log files.
• Burst filters can suppress event bursts based on burst size, duration, and other criteria (rather than the event
type). Burst filters use the EMS burst detection and suppression (BDS) feature.
HP NonStop Event Management System