Help us improve your experience.

Let us know what you think.

Do you have time for a two-minute survey?

Troubleshooting the QFX5220

 

QFX5220 Troubleshooting Resources Overview

To troubleshoot a QFX5220 problem, you can use:

  • Junos OS CLI

    The CLI is the primary tool for controlling and troubleshooting hardware, Junos OS, routing protocols, and network connectivity. CLI commands display information from routing tables, information specific to routing protocols, and information about network connectivity derived from the ping and traceroute utilities. For information about using the CLI to troubleshoot Junos OS, see the appropriate Junos OS configuration guide.

  • Alarms and LEDs on the network ports, management panel, and components

    When the Routing Engine detects an alarm condition, it lights the red or yellow alarm LED on the management panel as appropriate. In addition, you can also use component LEDs and network port LEDs to troubleshoot the QFX5220. For more information, see QFX5220 Management Panel.

  • JTAC

    If you need assistance during troubleshooting, you can contact the Juniper Networks Technical Assistance Center (JTAC) by using the Web or by telephone. If you encounter software problems, or problems with hardware components not discussed here, contact JTAC.

  • Knowledge Base articles–Knowledge Base.

QFX5220 Alarm Messages Overview

When a QFX5220 switch detects an alarm condition, it lights the red or yellow alarm LED on the management panel as appropriate. To view a more detailed description of the alarm cause, issue the show system alarms operational CLI command.

user@host> show system alarms

For thermal problems, the show chassis temperature-thresholds CLI command shows the cutoff temperatures for each level of alarm:

user@host> show chassis temperature-thresholds

It is also helpful to calculate the percentage of fan RPM, or duty cycle, use the following in root:

root@re0.~#i2cget -y -f 13 0x66 0x11

In this example, the system returned the hexadecimal value 0x07. Convert that value to decimal, which is 7 in this example. Then use this formula to get the duty cycle:

In this example, duty cycle = (7 +1)*6.25 = 50%

Chassis Alarm Messages

Chassis alarms indicate a failure on the device or one of its components. Chassis alarms are preset and cannot be modified.

Chassis alarms on QFX5220 devices have two severity levels:

  • Major (red)—Indicates a critical situation on the device that has resulted from one of the conditions described in Table 1. A red alarm condition requires immediate action.

  • Minor (yellow)—Indicates a noncritical condition on the device that, if left unchecked, might cause an interruption in service or degradation in performance. A yellow alarm condition requires monitoring or maintenance.

Table 1 describes the chassis alarm messages on QFX5220.

Junos OS Evolved systems, such as QFX5220, are based on a new alarm infrastructure, not all power supplies and fan alarms are supported. Table 1 shows these alarms.

Table 1: Chassis Alarm Messages for QFX5220

Component

Alarm Type

CLI Message

Recommended Action

Fans

Red (major)

Fan Tray fan-tray-number Absent

Install fan modules in the slots where they are absent.

Fan Tray fan-tray-number Failure

Remove and check fan module for obstructions. Reinsert the fan module. If the problem persists, replace the fan module.

sensor-location Temp Sensor Too Hot

Check the environmental conditions and alarms on other devices. Ensure that environmental factors (such as hot air blowing around the equipment) do not affect the temperature sensor. If the condition persists, the device might shut down.

Yellow (minor)

FAN fan-number Fan Sensor Fail

Remove and check fan module for obstructions. Reinsert the fan module. If the problem persists, check the system log for the message related to the sensor and report the message to customer service.

sensor-location Temp Sensor Too Warm

Check the environmental conditions and alarms on other devices. Ensure that environmental factors (such as hot air blowing around the equipment) do not affect the temperature sensor.

Power Supplies

Red (major)

PEM pem-number Not Powered

Install a power supply into the empty slot and ensure the power supply is powered.

Temperature sensors

Major (red)

FPC 0 Temperature Hot

Check environmental conditions and alarms on other devices. Ensure that environmental factors (such as hot air blowing around the equipment) do not affect the temperature sensor. if the condition persists, the device might shut down.

Minor (yellow)

FPC 0 Temperature Warm

Check environmental conditions and alarms on other devices. Ensure that environmental factors (such as hot air blowing around the equipment) do not affect the temperature sensor.

FPC 0 Temp Sensor Fail

Check the system log for the following error message and report the message to customer support:

Routing Engine

Major (red)

RE RE number /var partition is full

File storage is at capacity. Reduce unnecessary files to free space.

Minor (yellow)

RE RE number /var partition is high

File storage is reaching capacity. Reduce unnecessary files to free space.

Management Ethernet interface

Major (red)

Management interface management-interface-name down on node

Check whether a cable is connected to the management Ethernet interface, or whether the cable is defective. Replace the cable, if required.

Configuration Changes Leading to Unexpected QFX5220 Behavior

Port and channelization configuration procedures vary by release and can be the cause of unexpected port or switch behavior on QFX5220 switches. Ensure you are using the correct configuration and channelization procedures for your release. See Table 2.

Table 2: Release Dependant Port Configuration Changes

Model

Junos OS Evolved Release

Behavior

QFX5220-32C

19.1R1-EVO up to 20.1R1-EVO

If a single port is misconfigured, only the port that is misconfigured does not link.

20.1R1-EVO and later

If a single port is misconfigured, all ports return to default mode for the PIC. In the case of the QFX5220-32C, ports 0 to 31 default to 400 GbE and ports 32 and 33 default to 10 GbE. For example, suppose you had several ports with QSFP28 optics, manually configured for 100 GbE in ports 0-31. Later, you configured SFP+ port 33 also for 100 GbE. The software would allow the commit of the configuration, but the ports with the QSFP28 optics would default to 400 GbE , link would be down, and the SFP+ ports would default to 10 Gbps speed and the link would also be down.

QFX5220-128C

19.2R1-EVO up to 20.1R1-EVO

If a single port is misconfigured, only the port that is misconfigured does not link.

20.1R1-EVO and later

If a single port is misconfigured, all ports return to default mode for the PIC. In the case of the QFX5220-128C, ports 0 to 127 default to 100 GbE and ports 128 and 129 default to 10 GbE. For example, suppose you had several even numbered ports with QSFP+ optics, manually configured for 100 GbE in ports 0-31. Later, you configured SFP+ port 33 also for 40 GbE. The software would allow the commit of the configuration, but the ports with the QSFP+ optics would default to 100 GbE, link would be down, and the SFP+ ports would default to 10 Gbps speed and the link would also be down.

QFX5220-128C

19.2R1-EVO up to 20.2R1-EVO

You can configure any even-numbered QSFP28 port 40 Gbps speed. The system configures the next (odd) port as unused.

20.2R1-EVO and later

You can configure all 128 QSFP28 ports for 40 Gbps speed.

How to Troubleshoot QFX5220-128C Port Configuration Problems

Problem

Description: For Junos OS Evolved releases up to 20.2R1, if you try to configure an even port as 40 Gbps, or if you don’t follow the steps in the procedure in order, the link won’t come up. If you see the error message below in /var/log/messages the error is likely due to misconfiguration:

You can correct the configuration by following these steps:

  1. Delete the port speed configuration and commit the configuration:
    # delete chassis fpc 0 pic 0 port port-number speed 40g
    # commit
  2. Delete the unused port configuration and commit the configuration:
    # delete chassis fpc 0 pic 0 port port-number+1 unused
    # commit
  3. Verify the desired port can be configured as 40 Gbps (see QFX5220-128C Valid 100-Gbps Configurations) and perform the steps described in Port Configurations.

How to Troubleshoot QFX5220-128C Channelization Problems

Problem

Description: If you try to configure an ineligible port as channelized, or if you don’t follow the steps in the procedure in order, the link won’t come up, or only one sub-port interface is created. For example, in the output below from show chassis interfaces, there is only one sub-port created on et-0/0/20:

The configuration can be corrected by following these steps:

  1. Block the three ports following the misconfigured channelized port as unused, and commit the configuration if this step was not performed during the original configuration.
    # set chassis fpc 0 pic 0 port port-number+1 unused
    # set chassis fpc 0 pic 0 port port-number+2 unused
    # set chassis fpc 0 pic 0 port port-number+3 unused
    # commit
  2. Delete the channelization configuration and commit the configuration:
    # delete chassis fpc 0 pic 0 port port-number speed port-speed number-of-sub-ports 4
    # commit
  3. Delete the unused port configuration from the three ports folling the channelized port and commit the configuration:
    # delete chassis fpc 0 pic 0 port port-number+1 unused
    # delete chassis fpc 0 pic 0 port port-number+2 unused
    # delete chassis fpc 0 pic 0 port port-number+3 unused
    # commit
  4. Verify the desired port can be configured as channelized, see QFX5220-128C Valid 100-Gbps Configurations and perform the steps described in 4 x 25 Gbps Port Channelization.