15.3 Trouble Indicators

In this section we discuss various ways that you can gain clues to network problems from the Juniper Networks router itself. The router is designed in such a way as to provide real-time, accurate information on the status of the router's components , the status of network reachability, in some instances, and so on. Juniper Networks routers provide this information in three ways: system LEDs, the craft interface LCD display (on some models), and SNMP traps.

15.3.1 LEDs

As discussed in Chapter 3, each Juniper Networks router has a craft interface on the front of the router chassis. The craft interface contains the system LEDs that indicate the overall status of the router components. The LEDs shown in Figure 15-3 include the following:

• Alarm LEDs

• Routing engine LEDs

• FPC LEDs

Figure 15-3. M40 Craft Interface

The alarm panel shown in Figure 15-3 has a red alarm LED and a yellow alarm LED. Red alarms indicate a condition in which a service interruption could occur, such as total component failures. Yellow alarms are generally symptomatic of recoverable errors or maintenance alerts. If an alarm is present, the LCD of the craft interface (on certain models) will display more information about the cause of the system alarm.

Two alarm relay contacts allow external devices to be connected to the craft interface to provide an audible alarm, for example. An alarm cutoff button will silence the alarm, but will not clear the alarm. Resolving the issue causing the alarm is the only way to clear alarms.

The routing engine has two LEDs: OK and Fail. In a normal operational state, the OK LED will remain illuminated, solid and green. If this LED is blinking, it indicates that the routing engine is initializing. If the Fail LED is illuminated a solid red, it indicates that the routing engine has failed or that the system control board cannot make contact with the routing engine.

Over each number corresponding to an FPC slot there are also OK and Fail LEDs. In normal operational status, the OK LED will remain illuminated solid and green. If it is blinking, it indicates that the FPC is initializing. If the Fail LED should illuminate solid and red, it indicates that the router has lost contact with the FPC. This could indicate a hardware failure or an improperly inserted FPC.

Other LEDs are contained on the components themselves . For instance, the power supplies each have OK and Fail LEDs. There are instances in which a given router condition, such as increased temperature, can generate a red alarm and shut down one or both power supplies . Specific information about troubleshooting each hardware component of the Juniper Networks routers can be found in your hardware installation manual or online at www.juniper.net.

15.3.2 SNMP Traps

If you desire , SNMP can be enabled on the router. SNMP can then send traps, or alerts, to an NMS when certain conditions or events occur on the router. The NMS can be configured to alert network-engineering staff when a problem occurs. As discussed in Chapter 6, the following trap types are supported on all Juniper Networks routers:

• cold start/warm start

• Link up/link down

• Authentication failure

• jnxPowerSupplyFailure (enterprise)

• jnxChassisTraps (enterprise)

• jnxFanFailure (enterprise)

• jnxOverTemperature (enterprise)

• OSPF, MPLS, and BGP traps

If a power supply fails on New York in Figure 15-4, a jnxPowerSupplyFailure trap will be sent from New York to the NMS. If so configured, the NMS can then send a page (or other type of notification) to a network engineer immediately.

Figure 15-4. SNMP Traps

Using SNMP traps and an NMS should be part of a proactive network-management environment. By proactive, we mean that it is always better to monitor the health of the routers and the network overall constantly and consistently than to be surprised by a network outage . If the master power supply in New York fails, and a redundant power supply is available, the system will continue to operate by failing over to the backup power supply. If the master power supply fails and no one is alerted, the backup power supply could fail at some point and leave the router completely down. By using the SNMP traps, the initial problem can be resolved before a more major problem occurs.