Cpu Health Check; Viewing Cpu Health Check Results-Show Log Command - Extreme Networks ExtremeWare Version 7.8 Troubleshooting Manual

• If a health check checksum error message appears in the log, and the output of the
show diagnostics
use those two sources of information to determine the location of the problem.
• If backplane health check counts for missing or corrupted packets are increasing, but the log shows
no checksum error messages, the problem is probably a low-risk, transient problem—possibly a busy
CPU.
• If the log shows checksum error messages and the backplane health check counts for missing or
corrupted packets are increasing:
— Control data is probably being disrupted.
— The combination of the channel/slot information for the health check counters in the output of
the
show diagnostics
isolate the faulty module.
— Compare the backplane health check results to the results of the CPU health check. Because
backplane health check packets are sent out across the backplane data bus, but read back across
the same bus used by the CPU health check, packet errors might be occurring on the CPU control
path (slow path). In that case, user traffic might be largely unaffected, but protocol-level traffic
could be having problems.
• If the backplane health check shows no failures, but the log shows checksum error messages:
— If the checksum error messages occur infrequently, it might indicate a packet memory problem
that is being triggered sporadically; it might be a low-risk situation, but—if possible—you should
run the packet memory scan.
— If the checksum error messages occur frequently, user data is probably being affected; run the
packet memory scan as soon as possible.

CPU Health Check

The CPU health check routine in the system health checker tests the communication path between the
CPU and all I/O modules. The CPU health check uses five types of diagnostic packet. Those packets are
generated by the CPU on the I/O slots and sent back to the CPU through the CPU packet path.
Viewing CPU Health Check Results—show log Command
The CPU health check uses the same system log reporting mechanism as checksum validation, so you
can use the
show log
Log messages take the following form (date, time, and severity level have been omitted to focus on the
key information):
Sys-health-check [ type ] checksum error on < slot > prev= <0x m > cur= <0x n >
where type indicates the health check test packet type (INT, EXT, CPU), and slot indicates the probable
location of the error, from among the following:
• M-BRD—The main board of a Summit system.
• BPLANE—The backplane of an Alpine system.
• MSM-A, MSM-B, MSM-C, or MSM-D—The MSM modules of a BlackDiamond system.
• Slot n—The slot number for an I/O module in a BlackDiamond system.
Advanced System Diagnostics and Troubleshooting Guide
command shows excessive backplane health check error counts, you can usually
command and the checksum message information in the log can help
command to view health check status information.
System (CPU and Backplane) Health Check
79