Queuing and scheduling have the following prerequisites:

• You must be familiar with the “Using Modular QoS CLI” chapter.

• You are logged on to the switch.

• You are in the correct VDC. A VDC is a logical representation of a set of system resources. You can use the switchto vdc command with a VDC number.

Queuing and scheduling have the following configuration guidelines and limitations:

• If a no-drop class is paused and the IP traffic is received with the CoS value of the no-drop class, IP traffic is queued in default queue due to the dscp-to-queue mapping behaviour. This is applicable to Cisco Nexus 7700 Series switches by default. Note that the dscp-to-queue mapping can be disabled.

• Configure system-defined class maps with care because the changes occur immediately and traffic might be disrupted on all VDCs.

• Defining the CoS with the match cos command is not supported for custom configured class maps.

• Specifying DSCP values for a class map with the match dscp command is not supported for custom configured class maps.

• When you are working with 10-Gigabit Ethernet ports in the shared mode, the egress queuing policy applies to all the ports in the port group. With the 10-Gigabit Ethernet ports in shared mode, all the ports in the port group must be in the same VDC. For information about the shared and dedicated modes, see the Cisco Nexus 7000 Series NX-OS Interfaces Configuration Guide. For information about the port groups, see the Cisco Nexus 7000 Series Hardware Installation and Reference Guide.

• You cannot set either the queue limit or WRED on ingress 10-Gigabit Ethernet ports except for the 8 port, 10-Gigabit Ethernet I/O module.

• To ensure accurate hardware programming on Cisco M1 Series modules, when you add a physical interface to a port-channel, which already has a queuing policy applied, you must remove the queuing policy and reapply. Without this reapplication, queuing policy will not be correctly applied on all the interfaces. Otherwise, you must ensure that the queuing policy is applied on the port-channel interface only after all the physical ports are bundled into the port-channel.

• On Cisco M1 series modules, it may not be possible to configure actual values for traffic shaping. For example, on a 1 Gigabit interface with 65% average shaping, the output rate on the interface goes only up to 450Mbps, whereas with 70%, it goes to 850Mbps.

• When changing egress Class of Service (CoS) to queue mapping, ensure that you specify 2 or 3 seconds as the minimum time limit between changes. Otherwise, continuous traffic drop might occur.

• The Cisco M3 series modules do not support per-queue counters for egress drops (multicast, unknown unicast, or broadcasts). The egress drops will be per port and per Q-Default counter.

• Only 8e templates are supported on Cisco M3 series modules.

• The Cisco M3 series module supports only the network-qos template. This template contains all the CoS values that match the MTU size.

• All data traffic will be enqueued to the default queue of dot1q-tunnel port because this port is untrusted by default.

Queuing and scheduling are configured by creating policy maps of type queuing that you apply to either traffic direction of an interface. You can modify system-defined class maps, which are used in policy maps to define the classes of traffic to which you want to apply policies.

Additional considerations are as follows:

• Changes to system class maps take effect immediately across all VDCs.

  The specified CoS values immediately map to the new queues.

• Changes are disruptive.

  The traffic passing through ports of the specified port type experience a brief period of traffic loss. All ports of the specified type are affected. For example, if you change COS-to-queue mapping for the M1 10G egress interface type, all M1 10G ports in all VDCs experience a brief disruption.

• Performance can be impacted.

  If one or more ports of the specified type do not have a queuing policy applied that defines the behavior for the new queue, then the traffic mapping to that queue might experience performance degradation.

• If you change the CoS-to-queue mapping by modifying the queuing class maps, you must ensure that a new queuing policy was applied to all ports of that type that use the new queues.

• If you change the DSCP-to-ingress-queue mapping by modifying the queuing class maps, you must ensure that a new queuing policy is applied to all ports of that type that use the new queues.

• By default, nonused queues do not have an allocated buffer. Allocate buffers to these queues to avoid tail drop.

• Changes to system class-maps are made only on the default VDC.

For information about configuring policy maps and class maps, see “Using Modular QoS CLI.”

You can configure the congestion-avoidance features, which include tail drop and WRED, in any queue. You can configure one of the egress congestion management features, such as priority, shaping, and bandwidth, in output queues, and you can configure bandwidth in input queues.

We recommend that you modify the CoS value before you create a policy map. You can modify the CoS values that are matched by device-defined class map queues. You must assign each CoS value from 0 to 7 to one or more of the queues for each queue type. Each CoS value is used only once in each queue type.

We recommend that you modify the DSCP value before you create a policy map. You can modify the DSCP values that are matched by device-defined class map queues. You must assign each DSCP value from 0 to 63 to one or more of the queues for each queue type. Each DSCP value is used only once in each queue type.

The system-defined policy maps default-in-policy and default-out-policy are attached to all ports to which you do not apply a queuing policy map. The default policy maps cannot be configured. For more information about the default policy maps, see Table 5.

This example shows that if you downgrade from Release 4.0(3) to Release 4.0(2) and enter the show running-configuration command, the input default queuing policy has an unknown enum in the display:

switch# show running-config
version 4.0(2)
…
…
policy-map type queuing default-in-policy
class type queuing unknown enum 0
queue-limit percent 50
bandwidth percent 80
class type queuing unknown enum 0
queue-limit percent 50
bandwidth percent 20

If you copy and paste this configuration into any Cisco NX-OS release, the device sends errors while executing all the commands starting from the policy-map type queuing default-in-policy command. You can ignore these errors because they do not affect the performance of the device.

To configure queuing and scheduling, perform one of the following tasks:

For more information about the fields in the output from these commands, see the Cisco Nexus 7000 Series NX-OS Quality of Service Command Reference.

Note	Setting the ingress port CoS value makes the specified interfaces untrusted.

Note	Ensure that you are using the default queue for the port type that you are configuring. For information on the default queue for the port types, see "Using Modular QoS CLI.”

The following example shows how to configure ingress port CoS for 1-Gigabit Ethernet ports:

configure terminal
 policy-map type queuing untrusted_port_cos
  class type queuing 2q4t-in-q-default
   set cos 5
 interface ethernet 2/1
  service-policy type queuing input untrusted_port_cos

The following example shows how to configure ingress port CoS for 10-Gigabit Ethernet ports:

configure terminal
 policy-map type queuing untrusted_port_cos
  class type queuing 8q2t-in-q-default
   set cos 5
 interface ethernet 2/1
  service-policy type queuing input untrusted_port_cos

The following example shows how to configure the priority and queue limit features:

configure terminal
 class-map type queuing match-any 1p3q4t-out-pq1
  match cos 5-7
 class-map type queuing match-any 1p3q4t-out-q2
  match cos 3-4
 class-map type queuing match-any 1p3q4t-out-q3
  match cos 0-2
policy-map type queuing priority_queue1
 class type queue 1p3q4t-out-pq1
  priority
 class type queue 1p3q4t-out-q2
  bandwidth remaining percent 60
  queue-limit 1 mbytes
 class type queue 1p3q4t-out-q3
  bandwidth remaining percent 40
  queue-limit 2 mbytes

The following example shows how to configure the shaping and tail drop features:

configure terminal
 class-map type queuing match-any 1p3q4t-out-pq1
  match cos 5-7
 class-map type queuing match-any 1p3q4t-out-q2
  match cos 3-4
policy-map type queuing shape_dt
 class type queue 1p3q4t-out-pq1
  shape percent 50
  queue-limit cos 5 percent 10
  queue-limit cos 6 percent 10
 class type queue 1p3q4t-out-q2
  shape percent 25
  queue-limit cos 4 percent 15

Note	If the priority keyword is not specified for a pq1 queue, the queue is considered as a normal queue, not a priority queue.

The following example shows how to configure the bandwidth and WRED features for COS queues:

configure terminal
 class-map type queuing match-any 1p3q4t-out-pq1
  match cos 5-7
 class-map type queuing match-any 1p3q4t-out-q2
  match cos 3-4
policy-map type queuing bandwidth_wred
 class type queuing 1p3q4t-out-pq1
  bandwidth percent 50
  random-detect cos-based
  random-detect cos 5 minimum-threshold percent 10 maximum-threshold percent 30
  random-detect cos 6 minimum-threshold percent 40 maximum-threshold percent 60
 class type queuing 1p3q4t-out-q2
  bandwidth percent 25
  random-detect cos-based
  random-detect cos 4 minimum-threshold percent 20 maximum-threshold percent 40

The following example shows how to configure the bandwidth and WRED features for DSCP queues:

configure terminal
 class-map type queuing match-any 8q2t-in-q1
  match dscp 5-6
 class-map type queuing match-any 8q2t-in-q2
  match dscp 0-4
policy-map type queuing dscp_wred
 class type queuing 8q2t-in-q1
  bandwidth percent 50
  random-detect dscp-based
  random-detect dscp 5 minimum-threshold percent 10 maximum-threshold percent 30
  random-detect dscp 6 minimum-threshold percent 40 maximum-threshold percent 60
 class type queuing 8q2t-in-q2
  bandwidth percent 25
  random-detect dscp-based
  random-detect dscp 4 minimum-threshold percent 20 maximum-threshold percent 40

The following sample output from the show hardware qos dscp-to-queue ingress command displays the status of DSCP-to-queue mapping enabled in ingress direction on M-series modules:

Switch# show hardware qos dscp-to-queue ingress

status: Enabled
module_type : m-series

The table below summarizes the new and changed features for this document and shows the releases in which each feature is supported. Your software release might not support all the features in this document. For the latest caveats and feature information, see the Bug Search Tool at https://tools.cisco.com/bugsearch/ and the release notes for your software release.