Reference Guide

Usage
Information
The command output displays the storage device connected to each switch port and whether iSCSI
automatically detects it.
Example
OS10# show iscsi storage-devices
Interface Name Storage Device Name Auto Detected Status
-----------------------------------------------------------
ethernet1/1/23 EQL-MEM true
Supported
Releases
10.3.0E or later
Converged network DCB example
A converged data center network carries multiple traffic types (SAN, server, and LAN) that are sensitive to different aspects
of data transmission. For example, storage traffic is sensitive to packet loss, while server traffic is latency-sensitive. In a single
converged link, all traffic types coexist without imposing serious restrictions on others' performance. DCB allows iSCSI and
FCoE SAN traffic to co-exist with server and LAN traffic on the same network. DCB features reduce or avoid dropped frames,
retransmission, and network congestion.
DCB provides lossless transmission of FCoE and iSCSI storage traffic using:
● Separate traffic classes for the different service needs of network applications.
● PFC flow control to pause data transmission and avoid dropping packets during congestion.
● ETS bandwidth allocation to guarantee a percentage of shared bandwidth to bursty traffic, while allowing each traffic class
to exceed its allocated bandwidth if another traffic class is not using its share.
● DCBX discovery of peers, including parameter exchange (PFC, ETS, and other DCB settings), mismatch detection, and
remote configuration of DCB parameters.
● iSCSI application protocol TLV information in DCBX advertisements to communicate iSCSI support to peer ports
This example shows how to configure a sample DCB converged network in which:
● DCBx is enabled globally to ensure the exchange of DCBx, PFC, ETS, and ISCSI configurations between DCBx-enabled
devices.
● PFC is configured to ensure loseless traffic for dot1p priority 4, 5, 6, and 7 traffic.
● ETS allocates 30% bandwidth for dot1p priority 0, 1, 2, and 3 traffic and 70% bandwidth for priority 4, 5, 6, and 7 traffic.
● iSCSI is configured to use dot1p priority 6 for iSCSI traffic, and advertise priority 6 in iSCSI application TLVs.
● The default class-trust class map honors dot1p priorities in ingress flows and applies a 1-to-1 dot1p-to-qos-group and a
1-to-1 qos-group-to-queue mapping. In OS10, qos-group represents a traffic class used only for internal processing.
1. DCBX configuration (global)
Configure DCBX globally on a switch to enable the exchange of DCBX TLV messages with PFC, ETS, and iSCSI configurations.
OS10# configure terminal
OS10(config)# dcbx enable
2. PFC configuration (global)
PFC is enabled on traffic classes with dot1p 4, 5, 6, and 7 traffic. The traffic classes all use the default PFC pause settings for
shared buffer size and pause frames in ingress queue processing in the network-qos policy map. The pclass policy map honors
(trusts) all dot1p ingress traffic. The reserved class-trust class map is configured by default. Trust does not modify ingress
values in output flows.
OS10(config)# class-map type network-qos test4
OS10(config-cmap-nqos)# match qos-group 4
OS10(config-cmap-nqos)# exit
OS10(config)# class-map type network-qos test5
OS10(config-cmap-nqos)# match qos-group 5
OS10(config-cmap-nqos)# exit
OS10(config)# class-map type network-qos test6
OS10(config-cmap-nqos)# match qos-group 6
OS10(config-cmap-nqos)# exit
OS10(config)# class-map type network-qos test7
OS10(config-cmap-nqos)# match qos-group 7
OS10(config-cmap-nqos)# exit
OS10(config)# policy-map type network-qos test
732
Converged data center services