Troubleshooting Virtual Chassis Fabric

Problem

Cause

System logging and tracing operations place a load on the primary Routing Engine device in a VCF, taking processing cycles away from managing VCF operations. Logging in general, especially higher levels of logging and tracing operations, can have an impact on VCF stability.

Solution

To help ensure good convergence and stable operation in a large-scale VCF, system logging and tracing should always be used with discretion. During normal VCF operation, system logging should be set at or below the notice level, and tracing options disabled. When logging or tracing is necessary to troubleshoot a particular issue, use the following guidelines to minimize impact on VCF stability:

• Use the detail tracing option or system logging levels at or above the error level only for short periods of time during troubleshooting, and disable these settings after gathering enough information to begin analyzing the issue.

• Avoid logging the same level of information to more than one log file, which adds extra processing without the benefit of providing more information. Setting up logging to different files for different levels or facilities is a better option.

• Choose remote logging rather than local logging, and avoid logging to the console.

Problem

Cause

If one end of a link is configured as a VCP and the other is not configured as a VCP, the VCP link does not form.

The automatic VCP conversion feature, therefore, does not work in the following situations:

• a 40-Gbps QSFP+ or 10-Gbps SFP+ interface on one end of the link is already configured as a VCP.

  If you have previously removed a device from a VCF but haven’t used the request virtual-chassis vc-port delete command to convert the interface that was connected to the removed device out of VCP mode, the interface is still configured as a VCP.

  If you have removed a device from one Virtual Chassis or VCF and not changed the VCP port setting, the device being added to the VCF might also be configured as a VCP.

• a 40-Gbps QSFP+ port on an EX4300 switch, which is configured as a VCP by default, is interconnecting to a spine device.

Solution

Manually configure the interface that is not configured as a VCP into a VCP using the request virtual-chassis vc-port set command.

Problem

Solution

The device can assume the Routing Engine role for several minutes during setup before it receives the configuration from the primary Routing Engine, but eventually returns to the linecard role with no user intervention.