Congestion Avoidance Commands

bandwidth remaining

To specify how to allocate leftover bandwidth to various classes, use the bandwidth remaining command in policy map class configuration mode. To return to the system defaults, use the no form of this command.

bandwidth remaining [ratio ratio-value]

no bandwidth remaining [ratio ratio-value]

Syntax Description

ratio ratio-value

Specifies the amount of guaranteed bandwidth, based on a bandwidth ratio value. Range is 1 to 63 for main interfaces, and 1 to 255 for subinterfaces.

Command Default

No bandwidth is specified.

Command Modes

Policy map class configuration

Command History

Release

Modification

Release 7.0.12

This command was introduced.

Usage Guidelines

Bandwidth, bandwidth remaining, and shaping commands can be configured together in the same class.


Note


The bandwidth remaining command is supported only in the egress direction.


The available bandwidth is equally distributed among those queueing classes that do not have the remaining bandwidth explicitly configured.

The bandwidth remaining command is used to proportionally allocate bandwidth to the particular classes, but there is no reserved bandwidth capacity.

On egress, if the bandwidth remaining command is not present, then the bandwidth is shared equally among the configured queueing classes present in the policy-map.

Task ID

Task ID

Operations

qos

read, write

Examples

This example shows how the remaining bandwidth is shared by classes class1 and class2 in a 20:80 ratio.


RP/0/RP0/CPU0:router(config)#policy-map policy1
RP/0/RP0/CPU0:router(config-pmap)#class class1
RP/0/RP0/CPU0:router(config-pmap-c)#bandwidth remaining ratio 20
RP/0/RP0/CPU0:router(config-pmap-c)#exit
RP/0/RP0/CPU0:router(config-pmap)#class class2
RP/0/RP0/CPU0:router(config-pmap-c)#bandwidth remaining ratio 80
RP/0/RP0/CPU0:router(config-pmap-c)#exit
RP/0/RP0/CPU0:router(config-pmap)#exit
  

clear controller npu qos high-water-marks

To clear traffic class queue occupancy and delay high water marks data, use the clear controller npu qos high-water-marks command in XR EXEC mode.

clear controller npu qos high-water-marks [ monotonic | periodic ] interface type [ traffic-class number ] [ location ID ]

Syntax Description

monotonic

(Optional) Clears the monotonic high water marks. Monotonic high water marks are cleared if neither the monotonic or periodic keyword is used.

periodic

(Optional) Clears the periodic high water marks.

interface type

Interface type and number.

Specify an interface to clear high water marks for a single interface. Use the all keyword to clear high water marks for all interfaces.

traffic-class number

(Optional) Specify a traffic class number to clear high water marks for a single traffic class. High water marks are cleared for all traffic classes by default.

location ID

(Optional) Node ID.Specify a node id to clear high water marks for a single location.  Use the all keyword to clear high water marks for all locations.  All locations are cleared by default.

Command Modes

XR EXEC mode

Command History

Release Modification
Release 24.2.11

This command was introduced.

Usage Guidelines

The clear controller npu qos high-water-marks command form clears monotonic high water marks for all traffic classes on all interfaces. This is equivalent to the clear controller npu qos high-water-marks monotonic form of the command.

Use the interface, traffic-class and location keywords to limit the scope of the clear operation.

Use the periodic keyword to clear the periodic high water marks instead of the monotonic high water marks.

Examples

This command clears monotonic high water marks data for all traffic classes and all interfaces:

Router# clear controller npu qos high-water-marks interface all

This command clears monotonic high water marks data for traffic class 6 on interface fourHundredGigE 0/0/0/10:

Router# clear controller npu qos high-water-marks interface fourHundredGigE 0/0/0/10 traffic-class 6

This command clears monotonic high water marks data for all traffic classes and all interfaces:

Router# clear controller npu qos high-water-marks monotonic interface all

This command clears periodic high water marks data for all traffic classes on all interfaces on location 0/0/CPU0:

Router# clear controller npu qos high-water-marks periodic interface all location 0/0/CPU0 

This command clears periodic high water marks data for traffic class 3 on interface FH0/0/0/21:

Router# clear controller npu qos high-water-marks periodic interface FH0/0/0/21 traffic-class 3 

clear controller priority-flow-control statistics

To clear priority flow control statistics on an interface on a per-port or a per-traffic-class, per-port basis, use the clear controller priority-flow-control statistics command in XR EXEC mode.

clear controllers interface-type interfacepath-id priority-flow-control statistics traffic-class tc

Syntax Description

traffic class

tc —Traffic class to be cleared

Command Default

No default behavior or values

Command Modes

XR EXEC mode

Command History

Release Modification
Release 7.0.11

This command was introduced.

Usage Guidelines

None

Examples

This example shows a sample for the clear controllers priority-flow-control statistics command to clear statistics for traffic class 3:

Router#clear controller FourHundredGigE0/0/0/0 priority-flow-control statistics traffic-class 3

clear controller priority-flow-control watchdog statistics

To clear priority flow control watchdog statistics on an interface on a per-port or a per-traffic-class, per-port basis, use clear controller priority-flow-control watchdog-stats command in XR EXEC mode.

clear controllers interface-type interfacepath-id priority-flow-control watchdog-stats traffic-class tc

Syntax Description

traffic class

tc —Traffic class to be cleared

Command Modes

XR EXEC mode

Command History

Release Modification
Release 7.0.11

This command was introduced.

Usage Guidelines

None

Examples

This example shows how to clear priority flow control watchdog statistics for traffic class 3 on a FourHundredGigE 0/0/0/0 interface.


Router#clear controller FourHundredGigE 0/0/0/0 priority-flow-control watchdog-stats traffic-class 3

hw-module profile npu memory buffer-extended bandwidth-congestion-detection enable

To enable the ability to detect High Bandwidth Memory (HBM) congestion when you configure PFC in the buffer-extended mode, use the hw-module profile npu memory buffer-extended bandwidth-congestion-detection enable command in the XR Config mode.

To disable this feature, use the no form of this command.

hw-module profile npu memory buffer-extended location lc bandwidth-congestion-detection enable

Syntax Description

location

lc —Line card location

Command Default

This feature is disabled by default.

Command Modes

XR Config mode

Command History

Release Modification
Release 7.5.3

This command was introduced.

Usage Guidelines

No specific guidelines impact the use of this command.

Task ID

Task ID Operation
qos read, write

Examples

The following example shows how to enable the detection of HBM congestion when packets are egressing from an interface.

Router#config
Router(config)#hw-module profile npu buffer-extended location 0/6/CPU0 bandwidth-congestion-detection enable
Router(config)#commit
Router(config)#exit

hw-module profile npu memory buffer-extended bandwidth-congestion-protect enable

To enable triggering of global pause frames (X-off) whenever there's HBM congestion in the buffer-extended mode, use the hw-module profile npu memory buffer-extended bandwidth-congestion-protect enable command in the XR Config mode.

To disable this feature, use the no form of this command.

hw-module profile npu memory buffer-extended location lc bandwidth-congestion-protect enable

Syntax Description

location

lc —Line card location

Command Default

This feature is disabled by default.

Command Modes

XR Config mode

Command History

Release Modification
Release 7.5.4

This command was introduced.

Usage Guidelines

Configuring the hw-module profile npu memory buffer-extended bandwidth-congestion-protect enable command for line cards where you've configured headroom values exceeding 6144000 bytes could result in a commit error or the feature not being enabled.

You must reload the line card for the hw-module profile npu memory buffer-extended bandwidth-congestion-protect enable command to take effect.

Task ID

Task ID Operation
qos read, write

Examples

The following example shows how to enable the triggering of global pause frames (X-off) whenever there's HBM congestion.

Router#config
Router(config)#hw-module profile npu buffer-extended location 0/1/CPU0 bandwidth-congestion-protect enable
Router(config)#commit

hw-module profile priority-flow-control

To configure PFC threshold values, use the hw-module profile priority-flow-control command in XR Config mode. You can configure the values either in the buffer-internal mode or the buffer-extended mode. To return to the system defaults, use the no form of this command.

To disable this feature, use the no form of this command.

hw-module profile priority-flow-control location lc [ buffer-extended traffic-class value pause-threshold value unit [ headroom value unit ]| buffer-internal traffic-class value pause-threshold value unit headroom value unit [ ecn value unit | max-threshold value unit | probability-percentage value unit ]]

For 88-LC0-36FH-M line cards:

hw-module profile priority-flow-control location <lc> [ buffer-extended traffic-class value pause-threshold value unit [ headroom value unit ]]

Syntax Description

location

<lc> —Line card location

buffer-extended

traffic-class

— value from 0 through 7

Sets PFC threshold values for long-reach PFC deployments. You can configure the following parameters using the buffer-extended option:

pause-threshold — pause-threshold for traffic class in ms or us and in KB or MB for 88-LC0-36FH-M line cards.

For 88-LC0-36FH-M line cards, also configure:

headroom —headroom for traffic class in KB or MB

buffer-internal

traffic-class

— value from 0 through 7

Sets PFC threshold values for short-reach PFC deployments. You can configure the following parameters using the buffer-internal option:

traffic-class — value from 0 through 7

pause-threshold — pause-threshold for traffic class in bytes

headroom — headroom for traffic class in bytes

ecn — Explicit Congestion Notification (ECN) threshold in bytes

From Release 7.3.6

max threshold — ECN maximum threshold value in bytes

probability-percentage — mark probability in percent

Command Default

No default behavior or values.

Command Modes

XR Config

Command History

Release

Modification

Release 7.0.11

This command was introduced.

Release 7.3.1

This command isn’t supported.

Release 7.3.15

This command is supported.

Release 7.3.16

The default PFC configuration functionality that autodetects PFC threshold values and configures them for the buffer-internal and buffer-extended modes, was introduced.

Release 7.3.6

The options to configure ECN maximum threshold value and mark probability were added.

Usage Guidelines

  • If you add a new traffic class and configure PFC threshold values for the first time on that traffic class, you must reload the line card.

  • If you want to use the functionality to autodetect the PFC threshold values (using the default PFC configuration), delete the existing PFC configuration and then run the requisite command. You needn't reload the line card.

Task ID

Task ID

Operations

qos

read, write

Examples

This example shows how to configure the PFC threshold values using the buffer-internal mode.


RP/0/RP0/CPU0:router(config)#hw-module profile priority-flow-control location 0/1/CPU0
RP/0/RP0/CPU0:router(config-pfc-loc)#buffer-internal traffic-class 3 pause-threshold 403200 bytes headroom 441600 bytes ecn 224640 bytes
RP/0/RP0/CPU0:router(config-pfc-loc)#buffer-internal traffic-class 4 pause-threshold 403200 bytes headroom 441600 bytes ecn 224640 bytes

RP/0/RP0/CPU0:router(config-pfc-loc)#commit

This example shows how to configure the PFC threshold values using the buffer-extended mode.


RP/0/RP0/CPU0:router(config)#hw-module profile priority-flow-control location 0/6/0/1
RP/0/RP0/CPU0:router(config-pfc-loc)#buffer-extended traffic-class 3 pause-threshold 10 ms
RP/0/RP0/CPU0:router(config-pfc-loc)#buffer-extended traffic-class 4 pause-threshold 10 ms
RP/0/RP0/CPU0:router(config-pfc-loc)#commit

The default PFC configuration autodetects PFC threshold values and configures them for the buffer-internal and buffer-extended modes.

Configure the traffic class values for the buffer-internal and buffer-extended modes in the hw-module profile priority-flow-control command. Depending on the line card, the functionality configures the default threshold values.

For example, for buffer-internal mode on a line card, configure:


RP/0/RP0/CPU0:router(config)#hw-module profile priority-flow-control location 0/0/CPU0
RP/0/RP0/CPU0:router(config-pfc-loc)#buffer-internal traffic-class 3
RP/0/RP0/CPU0:router(config-pfc-loc)#buffer-internal traffic-class 4
RP/0/RP0/CPU0:router(config-pfc-loc)#exit
RP/0/RP0/CPU0:router(config-pfc-loc)#commit

For buffer-extended mode on a line card, configure:


RP/0/RP0/CPU0:router(config)#hw-module profile priority-flow-control location 0/2/CPU0
RP/0/RP0/CPU0:router(config-pfc-loc)#buffer-extended traffic-class 2
RP/0/RP0/CPU0:router(config-pfc-loc)#buffer-extended traffic-class 1
RP/0/RP0/CPU0:router(config-pfc-loc)#exit
RP/0/RP0/CPU0:router(config-pfc-loc)#commit
The following example shows how to configure PFC in buffer-internal mode with custom values for all parameters including max-threshold and probability-percentage .
Router(config)#hw-module profile priority-flow-control location 0/6/01
Router(config-pfc-loc)#buffer-internal traffic-class 3 pause-threshold 1574400 bytes headroom 1651200 bytes ecn 629760 bytes max-threshold 1416960  bytes probability-percentage 50
Router(config-pfc-loc)#buffer-internal traffic-class 4 pause-threshold 1574400 bytes headroom 1651200 bytes ecn 629760 bytes max-threshold 1416960  bytes probability-percentage 50
Router(config-pfc-loc)#commit

hw-module profile qos high-water-marks

To enable traffic class queue high water marks monitoring for all egress interfaces, configure the hw-module profile qos high-water-marks command in XR configuration mode. To disable the feature, use the no form of the command.

hw-module profile qos high-water-marks

Syntax Description

This command has no keywords or arguments.

Command Default

Traffic class queue high water marks monitoring is disabled. 

Command Modes

XR config mode (config)

Command History

Release Modification
Release 24.2.11

This command was introduced.

Usage Guidelines

Perform a manual reload of the chassis or all line cards after enabling or disabling this feature.

Examples

This example shows how to enable traffic class queue high water marks monitoring for all egress interfaces:

Router(config)# hw-module profile qos high-water-marks 

A manual reload of the chassis or all line cards is required to enable/disable Traffic Class High Water Marks Monitoring

Router(config)# commit

hw-module profile qos voq-mode

To configure the VOQ mode value or the fair VOQ mode value, use the hw-module profile qos voq-mode command in XR Config mode. To disable this mode, use the no form of the command. You can configure the normal VOQ mode value from 8 (default value) to 4 or conversely from 4 to 8.

hw-module profile qos voq-mode [ 4 | 8 | fair-4 | fair-8 ]

Syntax Description

4

Indicates the normal VOQ mode to configure to 4.

8

Indicates the normal VOQ mode to configure to 8.

fair-4

Indicates the fair VOQ mode to configure to fair-4.

fair-8

Indicates the fair VOQ mode to configure to fair-8.

Command Default

The default normal VOQ mode value is 8.

Command Modes

XR Config mode

Command History

Release Modification
Release 7.0.12

This command was introduced.

Release 7.3.3

The option to configure fair VOQ mode was introduced.

Usage Guidelines

The default normal VOQ mode value is 8. To change the value to 4, you must:

  1. Configure hw-module profile qos voq-mode 4 and commit.

  2. Remove the queuing service-policy from all interfaces including main interface, subinterface, and bundle.

  3. Run reload location all to reload all nodes on your router.


Note


Because this mode supports a maximum of four VOQs, egress policy matches only on queues 7, 6, 5, and 0.


After you commit the hw-module configuration, you must reload the chassis immediately before proceeding with any other operation. Else, existing configurations on the chassis may be affected, leading to unexpected behavior.

Task ID

Task ID Operation
qos read, write

Examples

The following example shows how to enable VOQ mode value 4 on the router.

RP/0/RP0/CPU0:router#config
RP/0/RP0/CPU0:router(config)#hw-module profile
RP/0/RP0/CPU0:router(config)#hw-module profile qos
RP/0/RP0/CPU0:router(config)#hw-module profile voq-mode 4
RP/0/RP0/CPU0:router(config)#commit
RP/0/RP0/CPU0:router#reload location all

Examples

The following example shows how to configure fair-4 VOQ mode:

RP/0/RP0/CPU0:router#config
RP/0/RP0/CPU0:router(config)#hw-module profile
RP/0/RP0/CPU0:router(config)#hw-module profile qos
RP/0/RP0/CPU0:router(config)#hw-module profile voq-mode fair-4
RP/0/RP0/CPU0:router(config)#commit
RP/0/RP0/CPU0:router#reload location all

hw-module voq-watchdog cardshut disable

To prevent the line cards from being shut down upon detecting stuck VOQs, use the hw-module voq-watchdog cardshut disable command in the XR Config mode.

hw-module voq-watchdog cardshut disable

Syntax Description

This command has no arguments or keywords.

Command Default

This feature is enabled by default.

Command Modes

XR Config mode

Command History

Release Modification
Release 24.2.11

This command was introduced.

Usage Guidelines

No specific guidelines impact the use of this command.

Task ID

Task ID Operation
qos read, write

Examples

The following example shows how to disable the shutdown action on the line card upon detecting stuck VOQs.

Router#config
Router(config)#hw-module voq-watchdog cardshut disable
Router(config)#commit

hw-module voq-watchdog feature enable

To enable the Virtual Output Queue (VOQ) watchdog feature, use the hw-module voq-watchdog feature enable command in the XR Config mode.

To disable this feature, use the no form of this command.

hw-module voq-watchdog feature enable

Syntax Description

This command has no arguments or keywords.

Command Default

This feature is disabled by default.

Command Modes

XR Config mode

Command History

Release Modification
Release 24.2.11

This command was introduced.

Usage Guidelines

No specific guidelines impact the use of this command.

Task ID

Task ID Operation
qos read, write

Examples

The following example shows how to enable the VOQ watchdog feature on your router.

Router#config
Router(config)#hw-module voq-watchdog feature enable
Router(config)#commit
Router(config)#exit

queue-limit

To specify or modify the maximum number of packets the queue can hold for a class policy configured in a policy map for each port, use the queue-limit command in policy map class configuration mode. To remove the queue packet limit from a class, use the no form of this command.

queue-limit value [unit ] | percent <1-100> | discard-class value <0-1> unit

no queue-limit

Syntax Description

value

Maximum threshold for tail drop in bytes. Range is from 1 to 4294967295.

unit

(Optional) Units for the queue limit value. Values can be:

  • bytes —bytes

  • kbytes —kilobytes

  • mbytes —megabytes

  • ms —milliseconds

  • packets —packets (default)

  • us —microseconds

Note

 

When the specified units is packets, packets are assumed to be 256 bytes in size.

percent

(Optional) Allows you to specify queue limit thresholds as a percentage of the total buffer limit for each port. This makes your provisioning model simpler and makes it easier for you to adjust the queue burst limit, irrespective of the queue’s service rate.

discard-class

(Optional) Allows you to configure dual queue limit thresholds for high priority and low priority. This option was introduced in Release 7.0.12.

value

  • 0 —higher priority flow

  • 1 —lower priority flow

(Optional) Units for the queue limit value. Values can be:

  • bytes —bytes

  • kbytes —kilobytes

  • mbytes —megabytes

  • ms —milliseconds

  • packets —packets (default)

  • us —microseconds

Command Default

The default value is 6 milliseconds for all queues including the high-priority queues.

Command Modes

Policy map class configuration

Command History

Release

Modification

Release 7.0.12

This command was introduced.

Usage Guidelines

When configuring the queue-limit command, you must configure one of the following commands: priority , shape average , bandwidth or bandwidth remaining , except for the default class. The default value is 6 milliseconds for all queues including the high-priority queues.

The queue-limit command is supported only in the egress direction.

Packets satisfying the match criteria for a class accumulate in the queue reserved for the class until they are serviced by the scheduling mechanism. The queue-limit command defines the maximum threshold for a class. When that threshold is reached, enqueued packets to the class queue result in tail drop (packet drop). Tail drop is a congestion avoidance technique that drops packets when a virtual output queue is full, until congestion is eliminated.

Use the show qos interface command to display the queue limit and other QoS values.

Queue Limit Default Values

These default values are used when queue-limit is not configured in the class.

If QoS is configured and random-detect is configured when the queue-limit is configured in time units, the guaranteed service rate (for the non-priority class) or the interface rate (for the priority class) is used to compute the queue-limit.

Task ID

Task ID

Operations

qos

read, write

Examples

This example shows how to configure two queue limits, one for higher priority and the other for lower priority.:


RP/0/RP0/CPU0:router(config)#policy-map egress_pol_dql
RP/0/RP0/CPU0:router(config-pmap)# class tc7
RP/0/RP0/CPU0:router(config-pmap-c)#priority level 1
RP/0/RP0/CPU0:router(config-pmap-c)#queue-limit ?
<1-4294967295>  Max Threshold for tail drop
  discard-class   Discard Class based QLIMIT (upto 8 values)
  percent         Configure queue limit value in percentage

RP/0/RP0/CPU0:router(config-pmap-c)#queue-limit discard-class ?
<0-1>  Discard Class  value
RP/0/RP0/CPU0:router(config-pmap-c)#queue-limit discard-class 0 100 ?
bytes    Bytes
  kbytes   Kilobytes
  mbytes   Megabytes
  ms       Milliseconds
  packets  Packets (default)
  us       Microseconds
  <cr>
RP/0/RP0/CPU0:router(config-pmap-c)#queue-limit discard-class 0 100 mbytes
RP/0/RP0/CPU0:router(config-pmap-c)#queue-limit discard-class 1 50 mbytes
RP/0/RP0/CPU0:router(config-pmap-c)#commit
RP/0/RP0/CPU0:router(config-pmap-c)#end

random-detect

To enable random early detection (RED), use the random-detect command in policy map class configuration mode. To remove RED, use the no form of this command.

random-detect [ [ [ min-threshold value ] [ max-threshold value ] probability percentage probability value ] ] [ discard-class value ]

no random-detect

Syntax Description

min-threshold value

Minimum threshold in number of packets. The value range of this argument is from 0 through 1073741823 in packets.

Note

 

On your router, the threshold numbers are converted into units of bytes (minimum: 6144, maximum: 390070272) and enforced.

max-threshold value

Maximum threshold in number of packets. The value range of this argument is from the value of the min-threshold argument through 1073741823. When the queue length exceeds the maximum threshold, RED drops all packets with the specified discard class value.

Note

 

On your router, the threshold numbers are converted into units of bytes (minimum: 6144, maximum: 390070272) and enforced.

units

(Optional) Units for the threshold values. Values can be:

  • bytes —bytes

  • gbytes —gigabytes

  • kbytes —kilobytes

  • mbytes —megabytes

  • ms —milliseconds

  • packets —packets (default)

  • us —microseconds

probability percentage probability value

Configure WRED mark probability in percent. The value range of this argument is from 1 through 100.

discard-class value

Discard-class based RED (up to 2 values, which is 0 and 1).

Command Default

Default unit for max-threshold and min-threshold is packets .

Command Modes

Policy map class configuration

Command History

Release

Modification

Release 7.0.12

This command was introduced.

Release 7.3.16

The functionality to configure WRED mark probability in percent was introduced.

Usage Guidelines

The RED congestion avoidance technique takes advantage of the congestion control mechanism of TCP. By randomly dropping packets before periods of high congestion, RED tells the packet source to decrease its transmission rate. Assuming the packet source is using TCP, it decreases its transmission rate until all the packets reach their destination, indicating that the congestion is cleared. You can use RED as a way to cause TCP to slow transmission of packets. TCP not only pauses, but it also restarts quickly and adapts its transmission rate to the rate that the network can support.

RED distributes losses in time and maintains normally low queue depth while absorbing traffic bursts. When enabled on an interface, RED begins dropping packets when congestion occurs at a rate you select during configuration.

When time units are used, the guaranteed service rate is used to compute thresholds.

When the value of the units argument is packets, packets are assumed to be 256 bytes in size.

Task ID

Task ID

Operations

qos

read, write

Examples

This example shows how to enable RED using a minimum threshold value of 1000000 and a maximum threshold value of 2000000:


RP/0/RP0/CPU0:router(config)# policy-map policy1
RP/0/RP0/CPU0:router(config-pmap)# class class1
RP/0/RP0/CPU0:router(config-pmap-c)# random-detect 1000000 2000000
  

service-policy (interface)

To attach a policy map to an input interface or output interface to be used as the service policy for that interface, use the service-policy command in the appropriate configuration mode. To remove a service policy from an input or output interface, use the no form of the command.

service-policy {input | output} policy-map

no service-policy {input | output} policy-map

Syntax Description

input

Attaches the specified policy map to the input interface.

output

Attaches the specified policy map to the output interface.

policy-map

Name of a service policy map (created using the policy-map command) to be attached.

Command Default

No service policy is specified.

Command Modes

Interface configuration.

Command History

Release

Modification

Release 7.0.12

This command was introduced.

Usage Guidelines

You can attach a single policy map to one or more interfaces to specify the service policy for those interfaces. The class policies composing the policy map are then applied to packets that satisfy the class map match criteria for the class. To apply a new policy to an interface, you must remove the previous policy. A new policy cannot replace an existing policy.

Task ID

Task ID

Operations

qos

read, write

Examples

This example shows policy map policy2 applied to HundredGigE 0/0/0/0.

RP/0/RP0/CPU0:router(config)# class-map class2
RP/0/RP0/CPU0:router(config)# match precedence ipv4 2
RP/0/RP0/CPU0:router(config-cmap)# exit

RP/0/RP0/CPU0:router(config)# policy-map policy2
RP/0/RP0/CPU0:router(config-pmap)# class-map class2
RP/0/RP0/CPU0:router(config-pmap-c)# set precedence 3
RP/0/RP0/CPU0:router(config-pmap)# exit

RP/0/RP0/CPU0:router(config)# HundredGigE 0/0/0/0
RP/0/RP0/CPU0:router(config-if)# service-policy input policy2
  
This example shows policy map policy 1 applied to Bundle-Ether interface.

RP/0/RP0/CPU0:router(config)# interface Bundle-Ether1
RP/0/RP0/CPU0:router(config-if)# service-policy input policy1 
RP/0/RP0/CPU0:router(config-if)# exit

service-policy (policy map class)

To use a service policy as a QoS policy within a policy map (called a hierarchical service policy), use the service-policy command in policy map class configuration mode. To disable a particular service policy as a QoS policy within a policy map, use the no form of this command.

service-policy [type qos] policy-map-name

no service-policy [type qos] policy-map-name

Syntax Description

type qos

(Optional) Specifies a QoS service policy.

policy-map-name

Name of the predefined policy map to be used as a QoS policy. The name can be a maximum of 40 alphanumeric characters.

Command Default

No service policy is specified.

Type is QoS when not specified.

Command Modes

Policy map class configuration

Command History

Release

Modification

Release 7.0.12

This command was introduced.

Usage Guidelines

The service-policy (policy-map class) command creates hierarchical service policies in policy-map class configuration mode.

This command is different from the service-policy (interface) command used in interface configuration mode.

The child policy is the previously defined service policy that is being associated with the class default of the parent policy-map. The new service policy using the preexisting service policy is the parent policy.

The service-policy (policy-map class) command has this restriction:

  • The priority command can be used only in the child policy.

Task ID

Task ID

Operations

qos

read, write

Examples

This example shows how to create a hierarchical service policy in the service policy called parent:


RP/0/RP0/CPU0:router(config)# policy-map child
RP/0/RP0/CPU0:router(config-pmap)# class class1
RP/0/RP0/CPU0:router(config-pmap-c)# priority level 1
RP/0/RP0/CPU0:router(config-pmap-c)# exit
RP/0/RP0/CPU0:router(config-pmap)# exit
RP/0/RP0/CPU0:router(config)# policy-map parent
RP/0/RP0/CPU0:router(config-pmap)# class class-default
RP/0/RP0/CPU0:router(config-pmap-c)# shape average 10000000
RP/0/RP0/CPU0:router(config-pmap-c)# service-policy child

shape average

To shape traffic to the indicated bit rate according to the algorithm specified, use the shape average command in policy map class configuration mode. To remove traffic shaping, use the no form of this command.

shape average {percent percentage | rate [units] | per-thousand value | per-million value }

no shape average

Syntax Description

percent percentage

Specifies the interface bandwidth in percentage. Values can be from 1 to 100.

rate

Average shaping rate in the specified units. Values can be from 1 to 4294967295.

units

(Optional) Units for the bandwidth. Values can be:

  • bps —bits per second (default)

  • gbps —gigabits per second

  • kbps —kilobits per second

  • mbps —megabits per second

per-thousand value

Specifies shape rate as parts per thousand of the available bandwidth.

per-million value

Specifies shape rate as parts per million of the available bandwidth.

Command Default

units: bps

Command Modes

Policy map class configuration

Command History

Release

Modification

Release 7.0.12

This command was introduced.

Usage Guidelines

The shape average command is supported only in the egress direction.

When you use the shape average command, egress shaping is done at the Layer 1 level and includes the Layer 1 header in the rate calculation. If you have both shape and bandwidth configured for a class, ensure that the shape percent value is always greater than the percent value for bandwidth. For bundled interfaces, shape average can be configured only as a percentage.

The priority and shape average commands can be configured together in the same class.

Task ID

Task ID

Operations

qos

read, write

Examples

This example sets traffic shaping to 50 percent of the parent shaper rate milliseconds:


RP/0/RP0/CPU0:router(config)# policy-map policy1
RP/0/RP0/CPU0:router(config-pmap)# class class1
RP/0/RP0/CPU0:router(config-pmap-c)# shape average percent 50
  

This example shows how to set traffic shaping to 100000 kbps:


RP/0/RP0/CPU0:router(config)# policy-map policy1
RP/0/RP0/CPU0:router(config-pmap)# class class1
RP/0/RP0/CPU0:router(config-pmap-c)# shape average 100000 kbps 
  

show controllers

To view Priority Flow Control (PFC) statistics, including PFC watchdog statistics, use the show controllers command in the XR EXEC mode.

show controllers { interface priority-flow-control [ statistics | watchdog-stats ] | all priority-flow-control { statistics | watchdog-stats } location all }

Syntax Description

interface

Displays statistics for an interface.

all

Displays statistics for all interfaces.

priority-flow-control

Specifies PFC mechanism.

statistics

Displays PFC statistics for an interface or all interfaces.

(Optional) For a specific interface.

watchdog-stats

Displays PFC watchdog statistics for an interface or all interfaces.

(Optional) For a specific interface.

location all

Displays PFC statistics for all nodes.

Command Modes

XR EXEC mode

Command History

Release

Modification

Release 7.5.5

Release 24.2.11

The show controllers all priority-flow-control statistics location all command form was added to display PFC statistics for all interfaces.

The show controllers all priority-flow-control watchdog-stats location all command form was added to display PFC watchdog statistics for all interfaces.

Release 7.5.4

For the show controllers priority-flow-control watchdog-stats command output, separate counters were added for total dropped packets and dropped packets.

For the show controllers priority-flow-control statistics command, an additional counter for PFC Tx pause frames per traffic class was added.

Release 7.0.11

The show controllers priority-flow-control statistics and show controllers priority-flow-control watchdog-stats command forms were introduced.

Usage Guidelines

  • From Release 7.5.5 (and Release 24.2.11), you can view the PFC and PFC Watchdog statistics for all interfaces using the show controllers all priority-flow-control statistics location all and show controllers all priority-flow-control watchdog-stats location all commands, respectively.
  • Running the clear controller priority-flow-control watchdog-stats command also clears the counters for the corresponding interfaces that the show controllers all priority-flow-control watchdog-stats location all command displays.
  • Running the clear controller priority-flow-control statistics command also clears the counters for the corresponding interfaces that the show controllers all priority-flow-control statistics location all command displays.
  • You can view PFC statistics and PFC Watchdog statistics per interface using the show controllers priority-flow-control statistics and show controllers priority-flow-control watchdog-stats commands, respectively.
  • Use the clear controller priority-flow-control statistics and clear controller priority-flow-control watchdog-stats commands to clear PFC statistics and PFC watchdog statistics counters, respectively.

Examples

This example shows a sample output of the show controllers priority-flow-control statistics command:

Router# show controllers hundredGigE 0/0/0/0 priority-flow-control statistics 

Priority flow control information for interface HundredGigE0/0/0/0:
Priority Flow Control:
    Total Rx PFC Frames: 0
    Total Tx PFC Frames: 1764273
    Rx Data Frames Dropped: 0
    CoS  Status  Rx Frames   
    ---  ------  ----------  
      0    Off      0
      1    Off      0
      2    Off      0
      3    Off      0
      4    Off      0
      5    Off      0
      6    Off      0
      7    Off      0
      

Examples

This example shows the sample output with counters for PFC Tx frames per traffic class.

Router# show controllers hundredGigE 0/0/0/4  priority-flow-control statistics 

Priority flow control information for interface HundredGigE0/0/0/4:
Priority Flow Control:
    Total Rx PFC Frames: 0
    Total Tx PFC Frames: 4832680
    Rx Data Frames Dropped: 1442056 (possible overflow)
    CoS  Status  Rx Frames   Tx Frames
    ---  ------  ----------  ----------
      0  on             0           0
      1  on             0           0
      2  on             0           0
      3  on             0      2416374
      4  on             0      2416306
      5  on             0           0
      6  on             0           0
      7  on             0           0

Examples

This example shows the sample output of show controllers hundredGigE 0/0/0/0 priority-flow-control watchdog-stats command:

Router# show controllers hundredGigE 0/0/0/0 priority-flow-control watchdog-stats
 
Priority flow control information for interface HundredGigE0/0/0/0:
 
Priority flow control watchdog statistics:
SAR: Auto restore and shutdown
-------------------------------------------------------------------------------------------------
Traffic Class            :       0        1        2        3        4        5        6        7 
-------------------------------------------------------------------------------------------------
Watchdog Events          :       0        0        0        0        0        0        0        0 
Shutdown Events          :       0        0        0        0        0        0        0        0 
Auto Restore Events      :       0        0        0        0        0        0        0        0 
SAR Events               :       0        0        0        0        0        0        0        0 
SAR Instantaneous Events :       0        0        0        0        0        0        0        0 
Total Dropped Packets    :       0        0        0        0        0        0        0        0

From Release 7.5.4 onwards, the output includes separate counters for total dropped packets and dropped packets:

Router# show controllers hundredGigE 0/1/0/43 priority-flow-control watchdog-stats
 
Priority flow control information for interface HundredGigE0/1/0/43: 
 
Priority flow control watchdog statistics:
SAR: Auto restore and shutdown
-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-
Traffic Class            :       0        1        2         3         4        5        6        7
-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-\-
Watchdog Events          :       0        0        0         3         3        0        0        0
Shutdown Events          :       0        0        0         3         3        0        0        0
Auto Restore Events      :       0        0        0         3         3        0        0        0
SAR Events               :       0        0        0      3510      3510        0        0        0
SAR Instantaneous Events :       0        0        0      1172      1172        0        0        0
Total Dropped Packets    :       0        0        0 941505767 941488166        0        0        0
Dropped Packets          :       0        0        0 314855466 314887161        0        0        0
      

Disregard the SAR Events and SAR Instantaneous Events entries because those numbers have no bearing on your operations.

Examples

This example shows a sample output of the show controllers all priority-flow-control statistics location all command:

Router# show controllers all priority-flow-control statistics location all     
 
   Interface                TC    Rx PFC  TxPFC  RxDropped
   
   FourHundredGigE0/0/0/0    0      0       0      NA
   FourHundredGigE0/0/0/0    7      0       0      NA
   FourHundredGigE0/0/0/0  all      0       0      0
   FourHundredGigE0/0/0/1    0      0       0      NA
   FourHundredGigE0/0/0/1    7      0       0      NA
   FourHundredGigE0/0/0/1  all      0       0      0
   FourHundredGigE0/0/0/2    0      0       0      NA
   FourHundredGigE0/0/0/2    7      0       0      NA
   FourHundredGigE0/0/0/2  all      0       0      0
   FourHundredGigE0/0/0/3    0      0       0      NA      
      

Examples

This example shows the sample output of the show controllers all priority-flow-control watchdog-stats location all command:

Router# show controllers all priority-flow-control watchdog-stats location all 
 
Interface               TC  Watchdog  Shutdown Auto Restore  SAR     SAR Instantaneous  Total Dropped   Dropped
                            Events    Events   Events        Events  Events        
                          
FourHundredGigE0/0/0/0   0   0         0        0             0       0                   0               0
FourHundredGigE0/0/0/0   7   0         0        0             0       0                   0               0
FourHundredGigE0/0/0/1   0   0         0        0             0       0                   0               0
FourHundredGigE0/0/0/1   7   0         0        0             0       0                   0               0
FourHundredGigE0/0/0/2   0   0         0        0             0       0                   0               0
FourHundredGigE0/0/0/2   7   0         0        0             0       0                   0               0
FourHundredGigE0/0/0/3   0   0         0        0             0       0                   0               0
FourHundredGigE0/0/0/3   7   0         0        0             0       0                   0               0
FourHundredGigE0/0/0/4   0   0         0        0             0       0                   0               0
  

show controllers npu packet-memory

To display the High Bandwidth Memory (HBM) congestion state and the current current memory usage of Shared Memory System (SMS) and HBM, use the show controllers npu packet-memory command in the XR EXEC mode.

show controllers npu packet-memory [ all | congestion | interface type | usage {instance { number | all }| verbose} ]

Syntax Description

all

Displays all details, including memory usage and congestion .

congestion

Displays Open Forwarding Abstraction (OFA) hardware congestion.

interface type

Displays the name of the IEEE 802.3 interfaces.

For type, you can specify a specific interface or use all to specify all interfaces.

usage instance number

Displays information for the specific device instance number selected.

usage instance all

Displays information for all device instances

verbose

Displays information details, including timestamp in milliseconds.

Command Default

None

Command Modes

XR EXEC mode

Command History

Release

Modification

Release 7.5.5

Counters to view buffer availability for Shared Memory System (SMS) and High Bandwidth Memory (HBM) were added.

Release 7.5.3

This command was introduced.

Usage Guidelines

The show controllers npu packet-memory command is supported only when you:

  • configure Priority Flow Control in the buffer-extended mode. (See hw-module profile priority-flow-control .)

  • configure the hw-module profile npu memory buffer-extended bandwidth-congestion-detection enable command to detect HBM congestion.

For details on detecting HBM congestion, and detailed examples of show controllers npu packet-memory , see the Modular QoS Configuration Guide for Cisco 8000 Series Routers.

Task ID

Task ID

Operations

interface

read

cisco-support

read

Examples

From Release 7.5.5, you can also view details of the buffer available for SMS (Buff-int-free Min WM ) and HBM (Buff-ext-free Min WM ) when you run the show controllers npu packet-memory usage instance all location all command.
Router#show controller npu packet-memory usage instance all location all 
HW memory Information For Location: 0/6/CPU0
 
---------------------------------------------------------------------------------------------------------------------
Timestamp(msec)                  | Device | Buff-int | Buff-int | Buff-ext | Buff-ext | Buff-int-free | Buff-ext-free
                                            Usage      Max WM   | Usage    | Max WM   | Min WM        | Min WM
---------------------------------------------------------------------------------------------------------------------
Wed 2023-08-30 23:47:40.918 UTC       0       1518       6668      17154      17656     293394          982846
Wed 2023-08-30 23:47:41.918 UTC       0       1227       5631      16010      16427     293685          983990
Wed 2023-08-30 23:47:42.919 UTC       0       1398       8295      15041      15734     293514          984959
Wed 2023-08-30 23:47:43.919 UTC       0       1765       8892      14744      15678     293147          985256
Wed 2023-08-30 23:47:41.011 UTC       1      10380      12419      37532      38165     284532          962468
Wed 2023-08-30 23:47:42.011 UTC       1      10463      11977      37315      38326     284449          962685
Wed 2023-08-30 23:47:43.013 UTC       1       9145      12604      37714      38242     285767          962286
Wed 2023-08-30 23:47:44.013 UTC       1      10996      13272      37429      38051     283916          962571

Examples

Run the show controllers npu packet-memory usage verbose instance all location all to view memory usage details in SMS and HBM.
RP/0/RP0/CPU0:Router#show controllers npu packet-memory usage verbose instance all location all 

HW memory Information For Location: 0/RP0/CPU0

* Option 'verbose' formatted data is for internal consumption.
--------------------------------------------------------------------
Timestamp(msec) | Device | Buff-int | Buff-int | Buff-ext | Buff-ext
                           Usage      Max WM   | Usage    | Max WM
--------------------------------------------------------------------
   1663958881006       0       2455       2676        637        640
   1663958882007       0       2461       2703        635        640
   1663958883007       0       2364       2690        635        640
   1663958884007       0      71603      75325       3183      18336
   1663958885008       0       2458       2852       1275       1279
   1663958886008       0       2484       2827       1275       1279

Run the show controllers npu packet-memory congestion verbose instance all location all command to view congestion details in SMS and HBM.

Router#show controllers npu packet-memory congestion verbose instance all location all
HW memory Information For Location: 0/RP0/CPU0
* Option 'verbose' formatted data is for internal consumption.
-------------------------------------------------------------------------------------------------------------------------------------
Timestamp(msec) |      Event | Device | Slice |    VOQ |VOQ-buff| Evicted-buff | Buff-int | Buff-int | Buff-int | Buff-ext | Buff-ext
                |      Type  |                         |int-WM  | int-WM       | UC-WM    | Usage    | Max WM   | Usage    | Max WM
-------------------------------------------------------------------------------------------------------------------------------------
1663958550349          0         0        5      534     16011        63969        65451      70410      70410      34405     34405
1663958551101          1         0        5      534         0            0          900       2440       2440          0         0
1663958557354          0         0        5      534     16011        63984        65493      70573      70573      34408     34408
1663958558354          1         0        5      534         0            0          915       2455       2455          0         0
1663958564606          0         0        5      534     16011        64002        65520      70081      70081      34532     34532
1663958565356          1         0        5      534         0            0          915       2417       2417          0         0

show controllers npu packet-memory interface

To view source queue buffer usage information in the packet memory space and pause duration of traffic class queues on PFC buffer-extended mode-enabled routers, use the show controllers npu packet-memory interface command in the XR EXEC mode.

show controllers npu packet-memory interface type { { rx-pause-percent | tx-pause-percent } [ detail | five-minute | one-minute | verbose ] location ID | sq-buffer-usage location ID }

Syntax Description

interface type

Interface type and number.

For the type argument, you can specify an interface and its number, or use the all keyword to view data for all interfaces.

If you select all, only the last one-minute and five-minute average entries are displayed, and not all 120 records.

rx-pause-percent

Specifies the pause duration of the input queues on the receiving router.

tx-pause-percent

Specifies the pause duration of the output queues on the transmitting router.

detail

(Optional) Displays a maximum of 120 records, at a frequency of one record per 250-millisecond interval.

[five-minute|one-minute|verbose]

(Optional) Displays traffic class queue pause frame information for a 1-minute or 5-minute duration.

If you don't specify the 1-minute or 5-minute options, 30 records are displayed for each traffic class at a frequency of one record per second. If you want to view the time stamp in a raw mode, choose verbose.

sq-buffer-usage ID

Displays the source queue buffer usage information in the packet memory space.

location ID

Specifies the node location. For ID, you can specify a specific node in the rack/slot/module notation or use all to specify all nodes.

Command Modes

XR EXEC mode

Command History

Release

Modification

Release 24.1.1

The sq-buffer-usage keyword was added to the show controllers npu packet-memory interface command form.

Release 24.2.11

The rx-pause-percent and tx-pause-percent keywords were added to the show controllers npu packet-memory interface command form

Usage Guidelines

Displaying the source queue buffer usage in all ports at all locations is not supported.

When you use the show controllers npu packet-memory interface type {rx-pause-percent|tx-pause-percent} location ID command form, 30 records are displayed for each traffic class at a frequency of one record per second. You can append the verbose option to this command form to display time stamps in a raw mode.

Examples

This example displays periodic information about the source queue buffer usage per port in the packet memory space

Router#show controllers npu packet-memory interface FourHundredGigE 0/0/0/23 sq-buffer-usage location 0/RP0/CPU0
Thu Feb  1 16:52:48.101 UTC

-------------------------------------------------------------------------
Node ID: 0/RP0/CPU0
Source Queue Buffer Usage Info for interface(s) FourHundredGigE0_0_0_23

Intf          Intf        TC      buffer-usage      congestion-state
name          handle
              (hex)
-------------------------------------------------------------------------
FH0/0/0/23   0x0f0001c8     3              0                 NO
FH0/0/0/23   0x0f0001c8     6              0                 NO

This example displays periodic information about source queue buffer usage in all ports in the packet memory space.

RP/0/RP0/CPU0:ios#show controllers npu packet-memory interface all sq-buffer-usage location 0/RP0/CPU0
Thu Feb  1 17:04:40.472 UTC

-------------------------------------------------------------------------
Node ID: 0/RP0/CPU0
Source Queue Buffer Usage Info for interface(s) all

Intf          Intf        TC      buffer-usage      congestion-state
name          handle
              (hex)
-------------------------------------------------------------------------
Hu0/0/0/0    0x0f000130     3              0                 NO
Hu0/0/0/0    0x0f000130     6              0                 NO
Hu0/0/0/35   0x0f000138     3              0                 NO
Hu0/0/0/35   0x0f000138     6              0                 NO
Hu0/0/0/34   0x0f000140     3              0                 NO
Hu0/0/0/34   0x0f000140     6              0                 NO
Hu0/0/0/33   0x0f000148     3              0                 NO
Hu0/0/0/33   0x0f000148     6              0                 NO
Hu0/0/0/32   0x0f000150     3              0                 NO
Hu0/0/0/32   0x0f000150     6              0                 NO
Hu0/0/0/31   0x0f000158     3              0                 NO
Hu0/0/0/31   0x0f000158     6              0                 NO
Hu0/0/0/30   0x0f000160     3              0                 NO
Hu0/0/0/30   0x0f000160     6              0                 NO
Hu0/0/0/29   0x0f000168     3              0                 NO
Hu0/0/0/29   0x0f000168     6              0                 NO
Hu0/0/0/28   0x0f000170     3              0                 NO
Hu0/0/0/28   0x0f000170     6              0                 NO
Hu0/0/0/27   0x0f000178     3              0                 NO
Hu0/0/0/27   0x0f000178     6              0                 NO
Hu0/0/0/26   0x0f000180     3              0                 NO
Hu0/0/0/26   0x0f000180     6              0                 NO
Hu0/0/0/25   0x0f000188     3              0                 NO
Hu0/0/0/25   0x0f000188     6              0                 NO
Hu0/0/0/24   0x0f000190     3              0                 NO
Hu0/0/0/24   0x0f000190     6              0                 NO
Hu0/0/0/20   0x0f000198     3              0                 NO
Hu0/0/0/20   0x0f000198     6              0                 NO
Hu0/0/0/19   0x0f0001a0     3              0                 NO
Hu0/0/0/19   0x0f0001a0     6              0                 NO
Hu0/0/0/15   0x0f0001a8     3              0                 NO
Hu0/0/0/15   0x0f0001a8     6              0                 NO
Hu0/0/0/8    0x0f0001b0     3              0                 NO
Hu0/0/0/8    0x0f0001b0     6              0                 NO
Hu0/0/0/1    0x0f0001b8     3              0                 NO
Hu0/0/0/1    0x0f0001b8     6              0                 NO
FH0/0/0/2    0x0f0001c0     3              0                 NO

Displaying the source queue buffer usage in all ports at all locations is not supported.

This example throws the 'Operation not supported' error in an attempt to display source queue buffer usage information for all ports at all locations.

RP/0/RP0/CPU0:ios#show controllers npu packet-memory interface all sq-buffer-usage location all
Thu Feb  1 17:13:35.010 UTC
Operation not supported on this location: all

Examples

Traffic Class Queue Pause Duration Examples:

In the following examples, input queue and output queue samples are provided for each example:

Example 1: Average Pause Duration in the Last Minute

Input Queue on R3: Displays the average pause duration of input queues on node 0/6/CPU0 interfaces for the last one-minute duration. Traffic for the traffic class 2 queue on interfaces FH0/6/0/21 and FH0/6/0/23 is paused for more than 50% of the time.

R3# show controllers npu packet-memory interface all tx-pause-percent one-minute location 0/6/CPU0 

-------------------------------------------------------------------------
Node ID: 0/6/CPU0
Source Queue Pause Percentage Info for interface(s) all
Intf               TC          Pause-Percentage 
name                                            
-------------------------------------------------------------------------
  FH0/6/0/10       2           0.00000 
  FH0/6/0/11       2           0.00000 
  FH0/6/0/13       2           0.00000 
  FH0/6/0/14       2           0.00000 
  FH0/6/0/15       2           0.00000 
  FH0/6/0/16       2           0.00000 
  FH0/6/0/18       2           0.00000 
  FH0/6/0/21       2          53.01604 
  FH0/6/0/22       2           0.00000 
  FH0/6/0/23       2          53.13991

Output Queue on R2: Displays the average pause duration of output queues on node 0/6/CPU0 interfaces for the last one-minute duration. Traffic for the traffic class 2 queue on interfaces FH0/6/0/21 and FH0/6/0/23 is paused for more than 50% of the time.

R2# show controllers npu packet-memory interface all rx-pause-percent one-minute location 0/6/CPU0 
-------------------------------------------------------------------------
Node ID: 0/6/CPU0
Out Queue Pause Percentage Info for interface(s) all
Intf               TC          Pause-Percentage 
name                                            
-------------------------------------------------------------------------
  FH0/6/0/10       2           0.00000 
  FH0/6/0/11       2           0.00000 
  FH0/6/0/13       2           0.00000 
  FH0/6/0/14       2           0.00000 
  FH0/6/0/15       2           0.00000 
  FH0/6/0/16       2           0.00000 
  FH0/6/0/18       2           0.00000 
  FH0/6/0/21       2          53.01604 
  FH0/6/0/22       2           0.00000 
  FH0/6/0/23       2          53.13991 

Example 2: Average Pause Duration for the Last Minute on One Interface

Input Queue on R3: Displays the average pause duration of input queues on node 0/6/CPU0 interface FourHundredGigE0_6_0_21 for the last one-minute duration. Traffic for the traffic class 2 queue on interface FourHundredGigE0_6_0_21 is paused for more than 50% of the time.

R3# show controllers npu packet-memory interface FH0/6/0/21 tx-pause-percent one-minute location 0/6/CPU0  
-------------------------------------------------------------------------
Node ID: 0/6/CPU0
Source Queue Pause Percentage Info for interface(s) FourHundredGigE0_6_0_21
Time                                  TC      Pause-Percentage 
-------------------------------------------------------------------------
Thu 2024-03-28 18:53:30.264 UTC       2       53.49027 

Output Queue on R2: Displays the average pause duration of output queues on node 0/6/CPU0 interface FourHundredGigE0_6_0_21 for the last one-minute duration. Traffic for the traffic class 2 queue on interface FH0/6/0/21 is paused for more than 50% of the time.

R2# show controllers npu packet-memory interface FH0/6/0/21 rx-pause-percent one-minute location 0/6/CPU0 

Node ID: 0/6/CPU0
Out Queue Pause Percentage Info for interface(s) FourHundredGigE0_6_0_21
Time                                  TC      Pause-Percentage 
-------------------------------------------------------------------------
Thu 2024-03-28 18:53:30.264 UTC       2       53.49027 
      

Example 3: Average Pause Duration for the Last 5 Minutes

Input Queue on R3: Displays the average pause duration of input queues on node 0/6/CPU0 interfaces for the last five minutes. Traffic for the traffic class 2 queue on interfaces FH0/6/0/21 and FH0/6/0/23 is paused for more than 50% of the time.

R3# show controllers npu packet-memory interface all  tx-pause-percent five-minute location 0/6/CPU0 
-------------------------------------------------------------------------
Node ID: 0/6/CPU0
Source Queue Pause Percentage Info for interface(s) all
Intf               TC          Pause-Percentage 
name                                            
-------------------------------------------------------------------------
  FH0/6/0/18       2           0.00000 
  FH0/6/0/21       2          53.24746 
  FH0/6/0/22       2           0.00000 
  FH0/6/0/23       2          53.23203 
  FH0/6/0/24       2           0.00000 
  FH0/6/0/30       2           0.00000 
  Hu0/6/0/17       2           0.00000 
  Hu0/6/0/20       2           0.00000 

Output Queue on R2: Displays the average pause duration of output queues on node 0/6/CPU0 interfaces for the last five minutes. Traffic for the traffic class 2 queue on interfaces FH0/6/0/21 and FH0/6/0/23 is paused for more than 50% of the time.

R2# show controllers npu packet-memory interface all rx-pause-percent five-minute location 0/6/CPU0 
-------------------------------------------------------------------------
Node ID: 0/6/CPU0
Out Queue Pause Percentage Info for interface(s) all
Intf               TC          Pause-Percentage 
name                                            
-------------------------------------------------------------------------
  FH0/6/0/10       2           0.00000 
  FH0/6/0/11       2           0.00000 
  FH0/6/0/13       2           0.00000 
  FH0/6/0/14       2           0.00000 
  FH0/6/0/15       2           0.00000 
  FH0/6/0/16       2           0.00000 
  FH0/6/0/18       2           0.00000 
  FH0/6/0/21       2          53.01604 
  FH0/6/0/22       2           0.00000 
  FH0/6/0/23       2          53.13991 
      

Example 4: Average Pause Duration for the Last 5 Minutes on One Interface

Input Queue on R3: Displays the average pause duration of input queues on interface FH0/6/0/21 on node 0/6/CPU0 for the last five minutes. Traffic for the traffic class 2 queue on interface FH0/6/0/21 is paused for more than 50% of the time.

R3# show controllers npu packet-memory interface FH0/6/0/21 tx-pause-percent five-minute location 0/6/CPU0 
        
-------------------------------------------------------------------------
Node ID: 0/6/CPU0
Source Queue Pause Percentage Info for interface(s) FourHundredGigE0_6_0_21
Time                                  TC      Pause-Percentage 
-------------------------------------------------------------------------
Thu 2024-03-28 18:53:00.242 UTC       2       53.31213 

Output Queue on R2: Displays the average pause duration of output queues on interface FH0/6/0/21 on node 0/6/CPU0 for the last five minutes. Traffic for the traffic class 2 queue on interface FH0/6/0/21 is paused for more than 50% of the time.

R2# show controllers npu packet-memory interface FH0/6/0/21 rx-pause-percent five-minute location 0/6/CPU0 
-------------------------------------------------------------------------
Node ID: 0/6/CPU0
Out Queue Pause Percentage Info for interface(s) FourHundredGigE0_6_0_21
Time                                  TC      Pause-Percentage 
-------------------------------------------------------------------------
                                         2       53.31213 
      

Example 5: Pause Duration for the Last 30 Seconds With a One-Second Interval

Input Queue on R3: Displays the pause duration of input queues on interface FH0/6/0/21. The traffic class 2 queue is paused for more than 50% of the time for the entire time duration.

R3# show controllers npu packet-memory interface FH0/6/0/21 tx-pause-percent location 0/6/CPU0        
-------------------------------------------------------------------------
Node ID: 0/6/CPU0
Source Queue Pause Percentage Info for interface(s) FourHundredGigE0_6_0_21
Time                                  TC      Pause-Percentage 
-------------------------------------------------------------------------
Thu 2024-03-28 18:53:31.265 UTC       2       53.57590 
Thu 2024-03-28 18:53:32.266 UTC       2       54.04276 
Thu 2024-03-28 18:53:33.267 UTC       2       53.99954 
Thu 2024-03-28 18:53:34.268 UTC       2       53.38932 
Thu 2024-03-28 18:53:35.269 UTC       2       54.52591 
Thu 2024-03-28 18:53:36.270 UTC       2       54.27677 
Thu 2024-03-28 18:53:37.271 UTC       2       52.78747 
Thu 2024-03-28 18:53:38.271 UTC       2       53.42725 
Thu 2024-03-28 18:53:39.271 UTC       2       53.34946 
Thu 2024-03-28 18:53:40.272 UTC       2       53.06505 
Thu 2024-03-28 18:53:41.272 UTC       2       53.50647 
Thu 2024-03-28 18:53:42.272 UTC       2       53.49165 
Thu 2024-03-28 18:53:43.272 UTC       2       52.52429 
Thu 2024-03-28 18:53:44.272 UTC       2       53.44903
..
..

Output Queue on R2: Displays the pause duration of output queues on interface FH0/6/0/21. The traffic class 2 queue is paused for more than 50% of the time for the entire time duration.

R2# show controllers npu packet-memory interface FH0/6/0/21 rx-pause-percent location 0/6/CPU0        
-------------------------------------------------------------------------
Node ID: 0/6/CPU0
Out Queue Pause Percentage Info for interface(s) FourHundredGigE0_6_0_21
Time                                  TC      Pause-Percentage 
-------------------------------------------------------------------------
Thu 2024-03-28 18:53:31.265 UTC       2       53.57590 
Thu 2024-03-28 18:53:32.266 UTC       2       54.04276 
Thu 2024-03-28 18:53:33.267 UTC       2       53.99954 
Thu 2024-03-28 18:53:34.268 UTC       2       53.38932 
Thu 2024-03-28 18:53:35.269 UTC       2       54.52591 
Thu 2024-03-28 18:53:36.270 UTC       2       54.27677 
Thu 2024-03-28 18:53:37.271 UTC       2       52.78747 
Thu 2024-03-28 18:53:38.271 UTC       2       53.42725 
Thu 2024-03-28 18:53:39.271 UTC       2       53.34946 
Thu 2024-03-28 18:53:40.272 UTC       2       53.06505 
Thu 2024-03-28 18:53:41.272 UTC       2       53.50647 
Thu 2024-03-28 18:53:42.272 UTC       2       53.49165 
Thu 2024-03-28 18:53:43.272 UTC       2       52.52429 
Thu 2024-03-28 18:53:44.272 UTC       2       53.44903
..
..

Example 6: Pause Duration for the Last 30 Seconds With a 250 ms Interval

Input Queue on R3: Displays the pause duration of the input queues on interface FH0/6/0/21. The traffic class 2 queue is paused for more than 50% of the time.

R3# show controllers npu packet-memory interface FH0/6/0/21 tx-pause-percent detail location 0/6/CPU0 
-------------------------------------------------------------------------
This CLI will fetch large number of records.
Integration with automation scripts is not recommended
-------------------------------------------------------------------------
Node ID: 0/6/CPU0
Source Queue Pause Percentage Info for interface(s) FourHundredGigE0_6_0_21
Time                                  TC      Pause-Percentage 
-------------------------------------------------------------------------
Thu 2024-03-28 18:53:30.514 UTC       2       52.22349 
Thu 2024-03-28 18:53:30.764 UTC       2       53.67103 
Thu 2024-03-28 18:53:31.014 UTC       2       53.63637 
Thu 2024-03-28 18:53:31.265 UTC       2       54.77273 
Thu 2024-03-28 18:53:31.515 UTC       2       53.72371 
Thu 2024-03-28 18:53:31.765 UTC       2       54.29221 
Thu 2024-03-28 18:53:32.015 UTC       2       53.04150 
Thu 2024-03-28 18:53:32.266 UTC       2       55.11364
..
..

Output Queue on R2: Displays the pause duration of the output queues on interface FH0/6/0/21. The traffic class 2 queue is paused for more than 50% of the time.

R2# show controllers npu packet-memory interface FH0/6/0/21 rx-pause-percent detail location 0/6/CPU0 
-------------------------------------------------------------------------
This CLI will fetch large number of records.
Integration with automation scripts is not recommended
-------------------------------------------------------------------------
Node ID: 0/6/CPU0
Out Queue Pause Percentage Info for interface(s) FourHundredGigE0_6_0_21
Time                                  TC      Pause-Percentage 
-------------------------------------------------------------------------
Thu 2024-03-28 18:53:30.514 UTC       2       52.22349 
Thu 2024-03-28 18:53:30.764 UTC       2       53.67103 
Thu 2024-03-28 18:53:31.014 UTC       2       53.63637 
Thu 2024-03-28 18:53:31.265 UTC       2       54.77273 
Thu 2024-03-28 18:53:31.515 UTC       2       53.72371 
Thu 2024-03-28 18:53:31.765 UTC       2       54.29221 
Thu 2024-03-28 18:53:32.015 UTC       2       53.04150 
Thu 2024-03-28 18:53:32.266 UTC       2       55.11364 

show controllers npu priority-flow-control

To display the current status and configured thresholds in a hardware module configuration, use show controllers npu priority-flow-control command in XR EXEC mode.

show controllers npu priority-flow-control <loc>

Command Default

No default behavior or values

Command Modes

XR EXEC mode

Command History

Release Modification
Release 7.5.4

The output added values for ECN maximum threshold value and mark probability.

Release 7.0.11

This command was introduced.

Usage Guidelines

None

Examples

This example shows sample output of show controllers npu priority-flow-control loc 0/0/CPU0 command:

RP/0/RP1/CPU0:router#show controllers npu priority-flow-control loc 0/0/CPU0 
Mon Oct 12 14:35:17.531 UTC
 
Location:    0/0/CPU0
PFC:          Enabled
TC    Pause-threshold    Resume-Threshold     Headroom
-------------------------------------------------------
3       403200 bytes        40320 bytes        441600 bytes
4       403200 bytes        40320 bytes        441600 bytes
RP/0/RP1/CPU0:router#

This example shows the sample output with custom values for all parameters including max-threshold and probability-percentage .

Router#show controllers npu priority-flow-control location all
 
Location:              0/6/CPU0
PFC:                    Enabled
PFC Mode:               buffer-internal
TC    Pause-threshold    Headroom       ECN            ECN-MAX         Prob-per
------------------------------------------------------------------------------
3    1574400 bytes      1651200 bytes   629760 bytes   1416960 bytes   50
4    1574400 bytes      1651200 bytes   629760 bytes   1416960 bytes   50

show controllers npu qos high-water-marks

To view traffic class queue occupancy and delay high water marks for each traffic class on an egress interface, use the show controllers npu qos high-water-marks command in the XR EXEC mode.

show controllers npu qos high-water-marks [ monotonic | periodic [ last number ] ] interface type [ traffic-class number ] [ location ID ]

Syntax Description

monotonic

(Optional) Displays monotonically increasing high water marks since bootup or the last clear operation.

Monotonic high water marks are displayed if neither the monotonic or periodic keyword is used.

periodic [last number]

(Optional) Displays per-interval high water marks from recent periodic data collection intervals.

Use the last number keyword-argument combination to view a specific number of collection intervals. The maximum number of intervals (6) are displayed by default.

interface type

Interface type and number. Specify an interface to view high water marks for a single interface.

Use the all keyword to view high water marks for all interfaces.

traffic-class number

(Optional) Specify a traffic class number to view high water marks for a single traffic class.

High water marks are displayed for all traffic classes by default.

location ID

(Optional) Node ID.

Specify a node id to view high water marks for a single location.

Use the all keyword to view high water marks for all locations. Data for all locations is displayed by default.

Command Modes

XR EXEC mode

Command History

Release

Modification

Release 24.2.11

The command was introduced.

Examples

Field Description for the show controllers npu qos high-water-marks command:

Table 1. Common fields
Field Description

Interval Start and End (periodic only)

The periodic collection interval number, and the start and end time of the interval.

TC_Number =

(Number range is 0-7)
Indicates the traffic class for the high water marks data displayed on that line. For periodic output, TC_ Number is only displayed for the traffic class's first periodic interval.
Table 2. Queue Occupancy fields
Field Description
Max Occupancy %

The maximum queue occupancy experienced by this traffic class as a percentage of the total queue size.

Due to limited queue quantization thresholds provided by the NPU, the max occupancy percentage and max occupancy kilobytes value below are an estimate of the actual maximum queue occupancy.

Max Occupancy kilobytes

The maximum queue occupancy experienced by this traffic class in kilobytes.

The kilobytes value is calculated with the assumption that all buffers are fully packed (i.e., all 384 bytes utilized in an SMS buffer). As a result, the displayed kilobytes value will be higher than the actual number of kilobytes queued in many cases.

Queue Delay ns

The delay in nanoseconds at the time the maximum queue occupancy high water mark occurred.

Src Sys Port

Slot/NPU/Slc/Gid

The Slot, NPU, Slice and GID identify the virtual output queue on which the queue occupancy high water mark occurred. The GID is the global identifier of the source system port whose packet was dequeued when the maximum queue occupancy high water mark occurred. 

In most cases, all ports on each slice share a virtual output queue.  Although the identified source system port sent the packet that was detected as the high water mark, other ports using the same virtual output queue may have also contributed to the burst of packets that caused the queue occupancy high water mark.

Note

 
When the fair-4 or fair-8 VOQ mode is configured, each source port has its own virtual output queue. When PFC buffer-internal mode is configured, each port shares a virtual output queue with other ports on the same slice interface group (IFG). There are two IFGs per slice.

Use the show controllers npu voq-usage command to see which other ports share a virtual output queue with the identified source system port.

Timestamp

(monotonic only)

The timestamp when the maximum queue occupancy high water mark was recorded. The timestamp corresponds to the time the high water mark information was read from the NPU, and not the timestamp when the high water mark was detected by the NPU. 

The NPU is queried for high water mark info every 30 seconds, so the timestamp indicates the end time of a 30-second timeframe that the high water mark occurred within.

For example, a timestamp of 16:56:44 indicates that the high water mark was observed sometime between 16:56:14 and 16:56:44.

Table 3. Queue Delay fields
Field Description
Max Queue Delay ns

The maximum delay experienced by this traffic class in nanoseconds.

Queue Occupancy %

The queue occupancy as a percentage of the total queue size at the time the maximum queue delay high water mark occurred.

Due to limited queue quantization thresholds provided by the NPU, the occupancy percentage and occupancy kilobytes value below are an estimate of the actual maximum queue occupancy.

Queue Occupancy kilobytes

The queue occupancy in kilobytes at the time the maximum queue delay high water mark occurred.

The kilobytes value is calculated with the assumption that all buffers are fully packed (i.e., all 384 bytes utilized in an SMS buffer). As a result, the displayed kilobytes value will be higher than the actual number of kilobytes queued in many cases.

Src Sys Port

Slot/NPU/Slc/Gid

The Slot, NPU, Slice and GID identify the virtual output queue on which the queue delay high water mark occurred. The GID is the global identifier of the source system port whose packet was dequeued when the maximum delay high water mark occurred. 

In most cases, all ports on each slice share a virtual output queue.  Although the identified source system port sent the packet that was detected as the high water mark, other ports using the same virtual output queue may have also contributed to the burst of packets that caused the maximum delay high water mark. 

Note

 
When the fair-4 or fair-8 VOQ mode is configured, each source port has its own virtual output queue. When PFC buffer-internal mode is configured, each port shares a virtual output queue with other ports on the same slice interface group (IFG). There are two IFGs per slice.

Use the show controllers npu voq-usage command to see which other ports share a virtual output queue with the identified source system port.

Timestamp (monotonic only)

The timestamp when the maximum delay high water mark was recorded. The timestamp corresponds to the time the high water mark information was read from the NPU, and not the timestamp when the high water mark was detected by the NPU.  

The NPU is queried for high water mark info every 30 seconds, so the timestamp indicates the end time of a 30-second timeframe that the high water mark occurred within.

For example, a timestamp of 16:56:44 indicates that the high water mark was observed sometime between 16:56:14 and 16:56:44.

Example 1: Monotonic High Water Marks for All Traffic Classes

The following output displays monotonic high water marks data for all traffic classes on interface fourHundredGigE 0/0/0/11:

Router# show controllers npu qos high-water-marks interface fourHundredGigE 0/0/0/11

Interface Name        =   FH0/0/0/11
Interface Handle      =        0x1F8
System Port Gid       =           96
Asic Instance         =            0
        
       Queue Occupancy High Water Marks                                Queue Delay High Water Marks
        
       Max Occupancy    Queue     Src Sys Port                         Max Queue   Occupancy     Src Sys Port
       %    kilobytes   Delay ns  Slot/NPU/Slc/Gid Timestamp           Delay ns    %  kilobytes  Slot/NPU/Slc/Gid Timestamp
---------------------------------------------------------------------------------------------------------------------------------- 
TC_0 =   6.00 30965     73728     0/0/2/40         04/08/23 08:39:35   102400     3.00  15482   0/0/1/44       04/05/23 12:22:05  
TC_1 =   0.00 0         0         0/0/0/0          -                   0          0.00  0       0/0/0/0        -                   
TC_2 =  25.00 129024    1114112   0/0/0/48         04/07/23 01:10:23   1179648    15.00 77414   0/0/0/48       04/07/23 21:40:53
TC_3 =  70.00 361267    8912896   0/1/1/56         04/02/23 08:41:44   8912896    70.00 361267  0/1/1/58       04/02/23 08:41:44
TC_4 =  40.00 206438    2228224   3/0/2/4          04/09/23 06:38:35   2359296    25.00 129024  3/0/2/5        04/04/23 18:30:56
TC_5 =   0.00 0         0         0/0/0/0          -                   0          0.00  0       0/0/0/0        -                   
TC_6 =  78.00 599       6437184   3/1/0/24         04/10/23 16:35:00   8628192    64.00 492     7/0/2/76       04/10/23 16:35:00   
TC_7 =  25.00 129024    139264    3/0/0/14         04/06/23 08:39:41   155648     15.00 77414   0/2/2/66       04/08/23 08:39:41

[ -----------     Occupancy High Water Marks     ---------------    ]  [ --------------   Delay High Water Marks  -----------  ]
        
      

Example 2: Monotonic High Water Marks for a Single Traffic Class

The following output displays monotonically increasing high water marks data (since bootup or the last clear operation) for traffic class 5 on interface fourHundredGigE 0/0/0/2.

Router# show controllers npu qos high-water-marks monotonic interface fourHundredGigE 0/0/0/2 traffic-class 5  

Interface Name        =    FH0/0/0/2
Interface Handle      =    0xF000120
System Port Gid       =            6
Asic Instance         =            0
        
       Queue Occupancy High Water Marks                                 Queue Delay High Water Marks
        
       Max Occupancy    Queue     Src Sys Port                          Max Queue  Occupancy     Src Sys Port
       %     kilobytes  Delay ns  Slot/NPU/Slc/Gid Timestamp            Delay ns   %  kilobytes  Slot/NPU/Slc/Gid Timestamp
------------------------------------------------------------------------------------------------------------------------------ 
TC_5 = 40.00 206438     6815744   3/0/0/15         11/11/23 17:43:30   1811939328  25.00 129024  7/1/2/89         11/27/23 11:21:26
        
[ -----------     Occupancy High Water Marks     ---------------    ]  [ --------------   Delay High Water Marks  ----------- ]
        
      

Example 3: Periodic High Water Marks for a Single Traffic Class

The following output displays high water marks data for the last three periodic collection intervals for traffic class 7 on interface fourHundredGigE 0/0/0//5.

Router# show controllers npu qos high-water-marks periodic last 3 interface fourHundredGigE 0/0/0/5 traffic-class 7 

Interface Name       =    FH0/0/0/5
Interface Handle     =    0xF000138
System Port Gid      =            9
Asic Instance        =            0                                                                                                        
        
                                             Queue Occupancy High Water Marks           Queue Delay High Water Marks 
        
                                             Max Occupancy   Queue    Src Sys Port      Max Queue  Occupancy       Src Sys Port 
       Interval Start      End               %     kilobytes Delay ns Slot/NPU/Slc/Gid  Delay ns   %     kilobytes Slot/NPU/Slc/Gid 
-----------------------------------------------------------------------------------------------------------------------------------
TC_7 = 1 12/01/23 17:46:30 12/01/23 17:46:59 50.00 258048    34680274 7/1/2/91          34680274   50.00 258048    7/1/2/91        
       2 12/01/23 17:45:58 12/01/23 17:46:30 60.00 309657    52296260 0/2/1/68          61348106   50.00 258048    7/1/2/91        
       3 12/01/23 17:45:30 12/01/23 17:45:58 40.00 206438    15290430 0/2/1/68          15290430   40.00 206438    0/2/1/68        
        
                                             [----- Occupancy High Water Marks ----- ]  [ ------- Delay High Water Marks ------  ]
        
      

Example 4: Periodic High Water Marks for All Traffic Classes

The following output displays periodic high water marks data for the last three periodic collection intervals for all traffic classes on interface fourHundredGigE 0/0/0/5:

Router# show controllers npu qos high-water-marks periodic last 3 interface fourHundredGigE 0/0/0/5 

Interface Name       =    FH0/0/0/5
Interface Handle     =    0xF000138
System Port Gid      =            9
Asic Instance        =            0
        
                                             Queue Occupancy High Water Marks           Queue Delay High Water Marks 
        
                                             Max Occupancy   Queue     Src Sys Port      Max Queue  Occupancy        Src Sys Port 
       Interval Start      End               %     kilobytes Delay ns  Slot/NPU/Slc/Gid  Delay ns   %      kilobytes Slot/NPU/Slc/Gid 
---------------------------------------------------------------------------------------------------------------------------------
TC_0 = 1 12/01/23 17:46:30 12/01/23 17:46:59 60.00 309657    60405506  0/1/0/60          81234030   50.00  258048    0/1/0/59
       2 12/01/23 17:45:58 12/01/23 17:46:30 80.00 412876    104607820 0/2/2/65          104607820  80.00  412876    0/2/2/67
       3 12/01/23 17:45:30 12/01/23 17:45:58 15.00 77414     9089660   0/1/0/6           11204567   6.00   30965     0/2/2/67

TC_1 = 1 12/01/23 17:46:30 12/01/23 17:46:59 0.05  258       288       3/2/0/37          512        0.05   258       3/2/0/38 
       2 12/01/23 17:45:58 12/01/23 17:46:30 0.05  258       288       3/2/0/39          258        0.05   258       3/2/0/37
       3 12/01/23 17:45:30 12/01/23 17:45:58 0.75  3870      512       3/2/0/39          768        0.50   2580      3/2/0/39
        
TC_2 = 1 12/01/23 17:46:30 12/01/23 17:46:59 50.00 258048    52046526  0/1/2/55          58048530   40.00  206438    0/1/2/55 
       2 12/01/23 17:45:58 12/01/23 17:46:30 40.00 206438    24690240  0/1/2/54          24690240   40.00  206438    0/1/2/54
       3 12/01/23 17:45:30 12/01/23 17:45:58 15.00 77414     6034820   7/2/0/111         8075420    6.00   30965     0/1/2/55        
                                             [----- Occupancy High Water Marks ----- ]  [ ------- Delay High Water Marks ------  ]
..
..    
TC_7 = 1 12/01/23 17:46:30 12/01/23 17:46:59 50.00 258048    34680274  7/1/2/91          34680274   50.00  258048    7/1/2/91
       2 12/01/23 17:45:58 12/01/23 17:46:30 60.00 309657    52296260  0/2/1/68          61348106   50.00  258048    7/1/2/91
       3 12/01/23 17:45:30 12/01/23 17:45:58 40.00 206438    15290430  0/2/1/68          15290430   40.00  206438    0/2/1/68        
                                             [----- Occupancy High Water Marks ----- ]  [ ------- Delay High Water Marks ------  ]
      

The following output displays periodic high water marks data for all traffic classes on interface fourHundredGigE 0/0/0/9:

Router# show controllers npu qos high-water-marks periodic interface fourHundredGigE 0/0/0/9

Interface Name       =    FH0/0/0/9
Interface Handle     =    0xF000158
System Port Gid      =           13
Asic Instance        =            0
        
                                             Queue Occupancy High Water Marks           Queue Delay High Water Marks 
        
                                             Max Occupancy   Queue     Src Sys Port      Max Queue  Occupancy        Src Sys Port 
       Interval Start      End               %     kilobytes Delay ns  Slot/NPU/Slc/Gid  Delay ns   %      kilobytes Slot/NPU/Slc/Gid 
---------------------------------------------------------------------------------------------------------------------------------

TC_0 = 1 12/01/23 17:45:58 12/01/23 17:46:30  6.00 30965     256       7/0/1/83          272        3.00   15482     7/0/1/83 
       2 12/01/23 17:45:30 12/01/23 17:45:58 25.00 129024    16777216  3/1/1/21          17825792   25.00  129024    7/0/1/83
       3 12/01/23 17:44:58 12/01/23 17:45:30 60.00 309657    268435456 7/2/2/103         285212672  50.00  361267    7/2/2/103
       4 12/01/23 17:44:31 12/01/23 17:44:58 15.00 77414     1048576   7/0/1/83          1114112    15.00  77414     7/0/1/83       
       5 12/01/23 17:43:59 12/01/23 17:44:31 15.00 77414     1068942   3/1/1/21          8912896    15.00  77414     3/1/1/21
       6 12/01/23 17:43:30 12/01/23 17:43:59 15.00 77414     6553602   0/1/1/59          6963204    15.00  77414     0/1/1/59
        
TC_1 = 1 12/01/23 17:45:58 12/01/23 17:46:30 25.00 129024    1114112   7/1/2/88          2228224    15.00 77414      7/1/2/88 
       2 12/01/23 17:45:30 12/01/23 17:45:58 60.00 309657    301989888 0/1/1/59          318767104  50.00 258048     0/1/1/59
       3 12/01/23 17:44:58 12/01/23 17:45:30 60.00 309657    188743688 7/1/2/88          188743688  60.00 309657     7/1/2/88
       4 12/01/23 17:44:31 12/01/23 17:44:58 25.00 129024    11796480  7/2/2/103         12451840   25.00 129024     7/1/2/88
       5 12/01/23 17:43:59 12/01/23 17:44:31 15.00 77414     9089660   0/1/1/59          11023450    6.00 30965      0/1/1/59
       6 12/01/23 17:43:30 12/01/23 17:43:59  6.00 30965     7372802   0/1/1/59          7782440     6.00 30965      7/1/2/88

TC_2 = 1 12/01/23 17:45:58 12/01/23 17:46:30  0.05 258       256       3/2/1/35          288         0.05 258        3/2/1/35
       2 12/01/23 17:45:30 12/01/23 17:45:58  0.25 1290      256       0/0/0/49          272         0.05 258        0/0/0/49
       3 12/01/23 17:44:58 12/01/23 17:45:30  0.75 3870      304       3/2/1/35          2202        0.50 2580       0/0/0/49
       4 12/01/23 17:44:31 12/01/23 17:44:58  0.75 3870      512       0/0/0/49          2506        0.50 2580       3/2/1/35
       5 12/01/23 17:43:59 12/01/23 17:44:31  3.00 15482     10412     3/2/1/35          10412       3.00 15482      3/2/1/35      
       6 12/01/23 17:43:30 12/01/23 17:43:59  3.00 15482     8192      0/0/0/50          10406       1.00 5160       0/0/0/48
        
                                              [--- Occupancy High Water Marks ---- ]  [ ---- Delay High Water Marks ---- ]
..
        
TC_7 = 1 12/01/23 17:45:58 12/01/23 17:46:30  6.00 30965     6578904   7/1/2/88          6578904     6.00 30965      7/1/2/88
       2 12/01/23 17:45:30 12/01/23 17:45:58 15.00 77414     5033164   3/0/2/6           5200936     15.00 77414     7/1/2/88
       3 12/01/23 17:44:58 12/01/23 17:45:30 50.00 258048    31457280  3/0/2/7           32505856    40.00 206438    3/0/2/6
       4 12/01/23 17:44:31 12/01/23 17:44:58 40.00 206438    19660800  7/1/2/90          20316160    40.00 206438    7/1/2/90
       5 12/01/23 17:43:59 12/01/23 17:44:31 15.00 77414     4560450   3/0/2/6           6004508     15.00 77414     3/0/2/6
       6 12/01/23 17:43:30 12/01/23 17:43:59 25.00 129024    6122880   7/1/2/88          7226976     15.00 77414     3/0/2/8

                                             [----- Occupancy High Water Marks ----]  [ ------- Delay High Water Marks ------ ]
      

show controllers npu voq in-extended-memory instance

To view the VOQs that are evicted to the HBM and the VOQs’ HBM buffer usage details, use the show controllers npu voq in-extended-memory instance command in the XR EXEC mode.

show controllers npu voq in-extended-memory instance { number | all } [ latest ] location ID

Syntax Description

number

Displays information for the specified device instance.

all

Displays information for all device instances.

latest

(Optional) Specifies that almost-instantaneous information be retrieved and displayed.

location ID

Displays information for the specified node.

For ID, specify a node in the rack/slot/module notation (0/0/CPU0, 0/RP0/CPU0 and so on) or use all to specify all nodes.

Command Modes

XR EXEC mode

Command History

Release

Modification

Release 24.2.11

This command was introduced.

Usage Guidelines

Important Caveat:

Do not use the show controllers npu voq in-extended-memory instance command in an automation script.

Other pointers:

  • If the PFC buffer-extended mode is enabled on a device, the associated VOQs are evicted to the HBM on priority. The remaining VOQs are retained in the Shared Memory Switch (SMS).
  • If PFC buffer-internal mode is enabled on a device, the associated VOQs are retained in the SMS and the remaining VOQs are evicted to the HBM.
  • If PFC isn’t enabled on a device, VOQs are evicted to the HBM based on the VOQs’ age and buffer usage.
  • With this feature, you can view up to 4000 records for these interfaces:
    • 400G and 100G (PFC buffer-internal and PFC buffer-extended modes)
    • 40G (PFC buffer-internal mode)

Examples

Output pointers for the show controllers npu voq in-extended-memory instance command:

  • Egress Interface–The egress interface of the virtual output queue.
  • VOQ_Base–Base VOQ ID
  • TC–Traffic Class number
  • Slice–Source slice number
  • Buff_Usage and In_Bytes–Buffer usage in blocks and in bytes

Example 1:  The following output displays VOQs that are evicted to the HBM for node 0/6/cpu0 and all instances. In this case, VOQs from device instance Device 1 are evicted to the HBM.

Router# show controllers npu voq in-extended-memory instance all location 0/6/CPU0 
        
* Use this CLI with caution.
* This should not be integrated with any automation scripts.
       
Total Entries  0  | Slot 6   | Device  0  | Percent in Evict Voq Buff(s) 0.000000             
---------------------------------------------------------------------------------
Egress Interface  | VOQ_Base  | TC  | Slice  |  Buff_Usage |  In_Bytes     |

Total Entries  2  | Slot 6   | Device  1  | Percent in Evict Voq Buff(s) 0.004883             
---------------------------------------------------------------------------------
Egress Interface  | VOQ_Base  | TC  | Slice  |  Buff_Usage  |  In_Bytes  |
---------------------------------------------------------------------------------
    FH0/6/0/0           28672       2     0         14722          87837459 
    FH0/6/0/3           28728       2     0         1049           6335369 

Total Entries  0  | Slot 6   | Device  2  | Percent in Evict Voq Buff(s) 0.000000             
---------------------------------------------------------------------------------
   Egress Interface  | VOQ_Base  | TC  | Slice  |  Buff_Usage  |  In_Bytes  |        

Example 2: The following output displays VOQs that are evicted to the HBM for node 0/6/cpu0 and all instances. In this case, the virtual output queue from device instance Device 0 is evicted to the HBM.

Router# show controllers npu voq in-extended-memory instance all location 0/6/CPU0  
        
* Use this CLI with caution.
* This should not be integrated with any automation scripts.
       
Total Entries  1  | Slot 6   | Device  0  | Percent in Evict Voq Buff(s) 0.002441             
---------------------------------------------------------------------------------
Egress Interface  | VOQ_Base  | TC  | Slice  |  Buff_Usage |  In_Bytes     |
---------------------------------------------------------------------------------
    FH0/6/0/22          28776       6     1         7140          46969050

Total Entries  0  | Slot 6   | Device  1  | Percent in Evict Voq Buff(s) 0.000000             
---------------------------------------------------------------------------------
Egress Interface  | VOQ_Base  | TC  | Slice  |  Buff_Usage  |  In_Bytes  |

Total Entries  0  | Slot 6   | Device  2  | Percent in Evict Voq Buff(s) 0.000000             
---------------------------------------------------------------------------------
   Egress Interface  | VOQ_Base  | TC  | Slice  |  Buff_Usage  |  In_Bytes  | 

show hw-module bandwidth-congestion-protect

To view the current status of global pause frames (X-off) in case of High Bandwidth Memory (HBM) congestion, use the show hw-module bandwidth-congestion-protect command in XR EXEC mode.

show hw-module bandwidth-congestion-protect location lc

Syntax Description

location

lc —Line card location

Command Default

No default behavior or values

Command Modes

XR EXEC mode

Command History

Release Modification
Release 7.5.4

This command was introduced.

Usage Guidelines

  • Use this command to view the status of the global pause frame (X-off) configuration after you configure the hw-module profile npu memory buffer-extended bandwidth-congestion-protect enable command.

  • The table lists the various possibilities for the command output based on your activity.

    If you...

    Configured field displays…

    Applied field displays…

    Action field displays…

    Configure the hw-module profile npu memory buffer-extended command

    Yes

    No

    Reload

    Use the no form of the hw-module profile npu memory buffer-extended command after configuring it, but before reloading the line card

    No

    No

    N/A

    Configure the hw-module profile npu memory buffer-extended command for a supported variant and reload the line card

    Yes

    Yes, Active

    Note

     

    Yes indicates that the configuration is programmed to the hardware, Active indicates that the global X-off functionality is active on the hardware.

    N/A

    Use the no form of the hw-module profile npu memory buffer-extended command when it is active, and commit the no form but don't reload the line card

    No

    Note

     

    At this stage, the output displays the user action and not the hardware status.

    No

    Note

     

    At this stage, the output displays the user action and not the hardware status.

    Reload

    Reload the line card after committing the no form of the hw-module profile npu memory buffer-extended command

    No

    Note

     

    At this stage, the output displays the hardware status.

    No

    Note

     

    At this stage, the output displays the hardware status.

    N/A

Task ID

Task ID Operation
qos read, write

Examples

This example shows how to view the current status of global pause frames (X-off) in case of High Bandwidth Memory (HBM) congestion using the show hw-module bandwidth-congestion-protect location 0/1/CPU0 command:

RP/0/RP1/CPU0:router#show hw-module bandwidth-congestion-protect location 0/1/CPU0 
 
Location   Configured     Applied          Action         
----------------------------------------------------------------------------

0/1/CPU0     Yes             No               Reload