Shelf Manager CLI
Fortinet Technologies Inc.
All events are logged, this includes normal system events as well as alarms and critical events. For example, insertion a
board generate a number of event log messages because the board is now present, has notified the shelf manager that
it’s ready to work, that it’s received the command to power up into operational mode, and that it’s now active in the
chassis. This is seen as clustered entries in the SEL where the same IPMC moves from M0-M1 all the way up to M3-M4,
where M4 is fully operating.
Alarm events such as fans starting to spin at slower rates due to age, or dips in the -48VDC power input or even blown
fuses are also added to the SEL.
The following example event log entry records that the rear fan in Fan Tray 1 (middle fan tray) is spinning below the
required RPM rate.
0x018D: Event: at Jan 1 00:02:15 1970; from:(0x10,0,0); sensor:(0x04,10); event:0x1(asserted):
"Upper Critical", Threshold: 0xff, Reading: 0xff
Where:
l
0x018D
The event ID, every new event increments by 1h. So the next event in the log file will be tagged with
0x018E. You can use this to keep track of what line you are on in the file.
l
Event: at Jan 1 00:02:15 1970
; The date and time that the event occurred.
l
from:(0x10,0,0)
; In this string, 0x10 is the IPMC controller to which the sensor is attached. In this message its
referencing an event monitored by the shelf manager in slot #1.
l
sensor:(0x04,10);
The first part 0x04 can be ignored, the 10 shows the ID of the Sensor that generated the
alarm.
l
event:0x1(asserted)
: Sensors show when an event is triggered because of something going wrong, but also
when they return to normal. In this case, asserted is the key point, meaning that the sensor has observed that the
device has gone outside the bounds or thresholds that have been set for it. Had this contained deasserted it would
mean the device has returned to its normal operating mode.
l
Upper Critical
", Threshold: ss a reference to the specific threshold level that the device exceeded. Some
events like thermal events could have multiple threshold levels. The first event could simply notify the shelf
manager that it’s getting too hot, to which the shelf manager would then notify the fans in the system to speed up
to help cool the board better, at which point the sensor would deassert the event. Another, higher level thermal
event could notify the shelf manager that the board has now exceeded its maximum thermal operating limit, to
which the shelf manager would tell it to deactivate and shut down.
l
Reading: 0xff
represents a raw value sent with the event, typically for more technical troubleshooting review
by the manufacturer. Data is rarely published for these values.
Reviewing the SEL messages from the IPMC, FRU and Sensor commands will show you exactly what devices are in
your chassis, what the FRU IDs are for the devices that connect to them, and all the sensor numbers that each of these
devices contain. You should be able to no identify the vast majority of both good and bad events that have happened to
your chassis and blades.
For more advanced users you can try using
–v
after many of the Shelf Manager commands or running the
sensordata
,
Threshold
, and
Fruinfo
commands to get more detailed information on each sensor and FRU
device.
Clearing SEL logs
Use the following command to clear SEL logs:
# clia sel clear
FortiGate-5144C 6.2.3 Chassis Guide
73
