Extreme Networks ExtremeWare XOS 11.1, Руководство пользователя

Страница: 127 / 496

System Health Checking
ExtremeWare XOS 11.1 Concepts Guide
127
The primary responsibility of the system health checker is to monitor and poll the ASIC error registers.
The system health checker processes, tracks, and reads the memory, parity, and checksum error counts.
The ASICs maintain counts of correctable and uncorrectable memory errors, as well as packets that
encountered checksum and parity errors. In a running system, some of these error counts may show
non-zero values. Occasional increments of these counters does not mean faulty hardware is detected or
that hardware requires replacement. If you see persistent increments of these counters, please contact
Extreme Networks Technical Support.
In addition, you can enable the system health checker to check the backplane, CPU, and I/O modules
by periodically sending diagnostic packets and checking the validity of the looped back diagnostic
packets.
In summary, two modes of health checking are available: polling and backplane diagnostic packets.
These methods are briefly described in the following:
●
Polling is always enabled on the system and occurs every 60 seconds by default. The system health
checker polls and tracks the ASIC counters that collect correctable and uncorrectable packet memory
errors, checksum errors, and parity errors on a per ASIC basis. By reading and processing the
registers, the system health check detects and associates faults to specific system ASICs.
●
Backplane diagnostic packets are disabled by default. If you enable this feature, the system health
checker tests the packet path for a specific I/O module every 6 seconds by default. The MSM sends
and receives diagnostic packets from the I/O module to determine the state and connectivity. (The
other I/O modules with backplane diagnostic packets disabled continue polling every 60 seconds by
default.)
System health check errors are reported to the syslog. If you see an error, please contact Extreme
Networks Technical Support.
Understanding the System Health Checker—Aspen 8810
Switch Only
On the Aspen 8810 switch, the system health checker tests the backplane, the CPUs on the MSM
modules, the I/O modules, the processes running on the switch, and the power supply controllers by
periodically forwarding packets and checking for the validity of the forwarded packets.
Two modes of health checking are available: polling (also known as control plane health checking) and
backplane diagnostic packets (also known as data plane health checking). These methods are briefly
described in the following:
●
Polling is always enabled on the system and occurs every 5 seconds by default. The polling value is
not a user-configured parameter. The system health check polls the control plane health between
MSMs and I/O modules, monitors memory levels on the I/O module, monitors the health of the
I/O module, and checks the health of applications and processes running on the I/O module. If the
system health checker detects an error, the health checker notifies the MSM.
●
Backplane diagnostic packets are disabled by default. If you enable this feature, the system health
checker tests the data link for a specific I/O module every 5 seconds by default. The MSM sends and
receives diagnostic packets from the I/O module to determine the state and connectivity. If you
disable backplane diagnostics, the system health checker stops sending backplane diagnostic packets.
System health check errors are reported to the syslog. If you see an error, please contact Extreme
Networks Technical Support.