The documentation set for this product strives to use bias-free language. For the purposes of this documentation set, bias-free is defined as language that does not imply discrimination based on age, disability, gender, racial identity, ethnic identity, sexual orientation, socioeconomic status, and intersectionality. Exceptions may be present in the documentation due to language that is hardcoded in the user interfaces of the product software, language used based on RFP documentation, or language that is used by a referenced third-party product. Learn more about how Cisco is using Inclusive Language.
This document provides helpful tips to view the initial set of hardware drop counters on a Nexus 5000 switch. This document is primarily for engineers who need immediate assistance in order to troubleshoot connectivity and/or performance problems on a Nexus 5000 switch.
Identify Packet Drops
Note: Gatos is a 50x0 ASIC. Carmel is a 55xx ASIC.
Determine if the end devices are in the process of flowcontrol. If so, enter the show interface flowcontrol and show interface priority-flow-control commands multiple times in order to check for any link-level pause. Frames or PPP frames are generated on the interfaces.
Enter the show hardware internal gatos/carmel event-history errorscommand. This shows you any errors recorded on Gatos or Carmel. If there are any pause frames in the network, check for oq_timestamp interrupts that are fired on Gatos. Carmel does not implement pause rate limits, so these interrupts are only shown for Gatos.
Determine if the MAC address is correctly recorded in the software and hardware for the end devices. If there is an issue with connectivity over the VLAN Switch Virtual Interface (SVI), then look for the respective MAC address on both Virtual Port Channel (vPC) peer switches in the hardware and software.
Enter these commands:
show mac address-table address
show platform fwm info hw-stm | inc <mac>
If the entry is not present in the hardware, then there is a problem. It causes a flood of MAC addresses on all ports and causes performance problems. It might also lead to connectivity issues.
Check for Queuing Drops
Enter the show queuing interface ethernet x/y command, and determine if there are ingress queuing discards:
Determine if span is enabled. Disable for span or rate-limit span if you span 10-gig ports to 1-gig, and follow the steps to mitigate from drops.
Determine if there is multicast traffic flow in the network. If there is, find more information about the multicast groups, flows, and fanouts. Implement multicast-optimize under the respective network Quality of Service (QoS) policy map, under system QoS, and perform a deep-dive WRT multicast troubleshoot.
If the interfaces are FEX Fabric Ports (FFP), then check the downstream FEX ports for any drops.
Check for Forwarding Manager (FwM) Drops
Enter the show platform fwm info pif Ethernet x/y | inc drop command. If there are drops, complete these steps:
Enter the show interface ethernet counters detailed (check for CRC/input/output for all the related interfaces) command.
Enter the show hardware internal gatos (or Carmel) port Ethernet x/y command and look for Capture Resource Center (CRC), Pause, and Frame errors.
Enter the show hardware internal gatos (or Carmel) port Ethernet x/y | inc inscommand. This gives you the ASIC ID and forwarding instance for the port.
Enter the show hardware internal gatos (or Carmel) asic x counters interruptcommand and look for any interrupt errors, such as CRC, len_err, mtu_vio, and the like.
Enter the show platform fwm info asic-errors <asic_id> command, and check for errors on the ASIC.
Enter the debug hardware internal carmel clear-counters interrupt / debug hardware internal carmel clear-counters error command in order to clear the Carmel counters.
N5K-2# debug hardware internal carmel clear? clear-counters Clear hardware counters for all carmel asic clear-interrupt Clear hardware interrupts for all carmel asic
Enter the show platform fwm info pif ethernet <> clear-stats verbose command in order to clear the Pif drops.
Enter the show platform fwm info l2mp myswid command. This command output is more specific for 55xx Carmel-based switches. Enter the command outputs on both vPC switches. If the myswid and peer swid are the same, then there is a problem. It might lead to traffic problems over the vPC peer link. Check for any known or new bugs in the bug database.
Contributed by Cisco Engineers
Vinayak Sudame, Colby Beam, and Somasundaram Jayaraman