Extreme Networks ExtremeWare Version 7.8, Troubleshooting Manual

Page: 33 / 122

Health Check Messages
Advanced System Diagnostics and Troubleshooting Guide 33
The following examples describe how these values apply to a BlackDiamond 6808:
• On a BlackDiamond 6808, if more than six fast-path errors are detected within one 20-second
window, a message is inserted into the system log. If this pattern recurs three times within three
windows, the system health check subsystem takes the action specified in the
configure sys-health-check
command.
• If fewer than six fast-path errors are detected within a single 20-second window, there is no
threshold violation, so no message is inserted in the system log.
• If more than six fast-path errors are detected within a single 20-second window, but no fast-path
errors are detected in other 20-second windows, an error message is inserted in the system log for
the fast-path window threshold violation, but no system health check action is taken.
NOTE
The state of the interpreting and reporting subsystem is configurable (enabled/disabled), as are the
values associated with the slow- and fast-path thresholds and bad window counters. However, these
attributes are currently accessible only under the instruction from Extreme Networks TAC personnel.
The default settings for these attributes have been found to work effectively under a broad range of
networking conditions and should not require changes.
Health Check Messages
As stated earlier, ExtremeWare maintains five types of system health check error counters, divided into
two categories: three slow path counters and two fast path counters.
• Slow path counters:
— CPU packet error—Data (control or learning) packet processed by the CPU and found to be
corrupted (a passive health check).
— CPU diagnostics error—CPU health check (an active health check)
—
Backplane diagnostics error—EDP diagnostics packets (an active health check)
• Fast path counters:
—
Internal MAC checksum errors
— External MAC checksum errors
Each of these system health check counters has an associated system log message type, to help focus
attention during troubleshooting. These message types are reported in the system log according to the
level of threat to the system. The message levels are:
• Alert messages
•
Corrective action messages
Alert Messages
These errors are inserted into the system log when the configured default error threshold is exceeded
within a given 20-second sampling window. When a threshold is exceeded, that window is marked as a
“bad” window and the interpreting and reporting subsystem inserts an error message into the system
log indicating the primary reason why the window was marked bad.