Brocade FastIron Ethernet Switch Administration Guide - Manual (Page 300)

Sysmon starts the timer based on the specified timer setting, with the default value as three minutes.
After the interval specified by the timer, the utility checks the hardware error registers. If the sysmon
utility detects an error in a hardware error register, it increments the relevant error count by 1.
Otherwise, it restarts the timer and waits for the given interval. Hardware error registers are cleared
when read, so after Sysmon reads the value, they are reset to zero.
Sysmon checks the value of the error counters it maintains and the values specified in the sysmon
threshold. If the value of the error counters exceeds the matching threshold, it takes the action
specified (logs internally or to the syslog). Otherwise, it restarts the timer and waits for the specified
interval before checking for errors again.
To ensure that logging repeating errors does not cause the logs to overflow, you can specify a back-off
value that allows the utility to skip the specified number of error instances before logging again. If the
error count is smaller than the specified log back-off value, the utility logs the error to the internal log or
syslog, restarts the timer and waits for the specified interval before checking for errors again.
Configuration notes and feature limitations
• While system monitoring is supported on all FastIron devices, the types of errors monitored vary
according to devices. On FSX devices, the sysmon utility monitors the following for errors:
‐ Fabric Adapter (FA) for processing and link errors.
‐ Cross Bar (XBAR) or Switch Fabric Module (SFM) for processing and link errors.
‐ Packet processor (PP) for link errors.
On FCX and ICX devices, the sysmon utility monitors the following errors:
• ‐ Link errors.
‐ ECC errors.
• By default, system monitoring starts on system boot up and runs in the background every three
minutes. You can configure, disable, or enable, the time interval through the CLI; however, if you
define the system monitoring interval at the global level, this value overrides the individual settings.
Valid range for the sysmon timer is 1 to 60 minutes.
• You can define a system monitoring threshold that is defined as N/W, where N is the number of
error events in a specified window (W) of consecutive polling periods. When the threshold is
reached, the action that is defined is performed. The threshold enables the sysmon utility to ignore
random errors that occur because of corrupted data coming in to the device, and perform the action
only for errors generated because of device failure. A threshold of 1/W means no threshold.
• You can choose the log action as either to the internal sysmon buffer or to the syslog. If you choose
the internal sysmon buffer, logs that are written beyond the limit of the sysmon buffer rolls over. On
the other hand, if you choose logging to syslog, messages are sent to the configured syslog
servers.
Configure system monitoring
You can use the following commands at the privileged EXEC level to globally configure the sysmon
utility:
• disable system-monitoring all on page 301
• enable system-monitoring all on page 301
• sysmon timer on page 301
In addition, you can enable or disable system monitoring for each event type from the CLI, with each
event type having separate threshold and log back off values.
Configuration notes and feature limitations
300 FastIron Ethernet Switch Administration Guide
53-1003075-02