The Event Tracer feature provides a binary trace facility for troubleshooting Cisco IOS software. This feature gives Cisco service representatives additional insight into the operation of the Cisco IOS software and can be useful in helping to diagnose problems in the unlikely event of an operating system malfunction or, in the case of redundant systems, route processor switchover.
This feature is intended for use as a software diagnostic tool and should be configured only under the direction of a Cisco Technical Assistance Center (TAC) representative.
Event tracing works by reading informational messages from specific Cisco IOS software subsystem components that have been preprogrammed to work with event tracing, and by logging messages from those components into system memory. Trace messages stored in memory can be displayed on the screen or saved to a file for later analysis.
By default, trace messages saved to a file are saved in binary format without applying additional processing or formatting. Saving messages in binary format allows event tracing to collect informational messages faster and for a longer time prior to a system malfunction or processor switchover. Optionally, event trace messages can be saved in ASCII format for additional file processing.
The Event Tracer feature can support multiple traces simultaneously. To do this, the feature assigns a unique ID number to each instance of a trace. This way, all messages associated with a single instance of a trace get the same ID number. Event tracing also applies a timestamp to each trace message, which aids in identifying the message sequence.
The number of trace messages stored in memory for each instance of a trace is configurable up to 65536 entries. As the number of trace messages stored in memory approaches the configured limit, the oldest entries are overwritten with new messages, which continues until the event trace is terminated.
Event tracing can be configured in "one-shot" mode. This is where the current contents of memory for a specified component are discarded and a new trace begins. New trace messages are collected until the message limit is reached, at which point the trace is automatically terminated.