Cisco Nexus 5500 Series Release Notes, Release 7.x
Release Date: January 29, 2014 Last Modified: March 02, 2015 Current Release: NX-OS Release 7.1(0)N1(1b)
This document describes the features, caveats, and limitations for the Cisco Nexus 5500 devices and the Cisco Nexus 2000 Series Fabric Extenders. Use this document in combination with documents listed in the “Related Documentation” section.
The Cisco NX-OS software is a data center-class operating system built with modularity, resiliency, and serviceability at its foundation. Based on the industry-proven Cisco MDS 9000 SAN-OS software, Cisco NX-OS helps ensure continuous availability and sets the standard for mission-critical data center environments. The highly modular design of Cisco NX-OS makes zero-effect operations a reality and enables exceptional operational flexibility.
Several new hardware and software features are introduced for the Cisco Nexus 5500 Series device and the Cisco Nexus 2000 Series Fabric Extender (FEX) to improve the performance, scalability, and management of the product line.
Cisco Nexus Devices
The Cisco Nexus devices include a family of line-rate, low-latency, lossless 10-Gigabit Ethernet, Cisco Data Center Ethernet, Fibre Channel over Ethernet (FCoE), and native Fibre Channel devices for data center applications.
For information about the Cisco Nexus 5500 Series, see the Cisco Nexus 5500 Series Platform Hardware Installation Guide.
Cisco Nexus 2000 Series Fabric Extenders
The Cisco Nexus 2000 Series Fabric Extender (FEX) is a highly scalable and flexible server networking solution that works with the Cisco Nexus 5500 Series devices to provide high-density and low-cost connectivity for server aggregation. Scaling across 1-Gigabit Ethernet, 10-Gigabit Ethernet, unified fabric, rack, and blade server environments, the FEX is designed to simplify data center architecture and operations.
The FEX integrates with its parent Cisco Nexus device, which allows zero-touch provisioning and automatic configuration. The FEX provides a single point of management that supports a large number of servers and hosts that can be configured with the same feature set as the parent Cisco Nexus 5500 Series switch, including security and quality of service (QoS) configuration parameters. Spanning Tree Protocol (STP) is not required between the Fabric Extender and its parent switch, because the Fabric Extender and its parent switch allow you to enable a large multi-path, loop-free, active-active topology.
Software is not included with the Fabric Extender. Cisco NX-OS software is automatically downloaded and upgraded from its parent switch. For information about configuring the Cisco Nexus 2000 FEX, see the “Configuring the Fabric Extender” chapter in the Cisco Nexus 5500 Series NX-OS Layer 2 Switching Configuration Guide, Release 7.x.
The Cisco NX-OS software supports the Cisco Nexus devices. Starting with Cisco NX-OS Release 7.0(0)N1(1), the Cisco Nexus 5010 and 5020 switches are not supported. You can find detailed information about supported hardware in the Cisco Nexus 5500 Series Hardware Installation Guide.
Table 2 shows the hardware supported by Cisco NX-OS Release 7.x software.
Table 2 Hardware Supported by Cisco NX-OS Release 7.x Software
BPDU Guard can be can be activated on disallowed edge trunk VLANs. This is done by configuring both sides of the link as either trunks or access interfaces.
CTS with FabricPath
The Cisco TrustSec security architecture has been extended to support Cisco FabricPath environments including those using VPC+. CTS packet classification can occur before or as traffic enters the fabric, at which point packet tags are preserved through the fabric for the purpose of applying security policy to the data path.
Dynamic ARP Inspection Enhancement
Dynamic ARP Inspection (DAI) can validate ARP packets against user-configured ARP access control lists (ACLs). DAI can be configured to drop ARP packets when the IP/MAC addresses in the packets are invalid. This is done by configuring ARP based ACLs.
IPv6 vPC/vPC+ Keepalive Support
IPv6 support for vPC/vPC+ provides IPv6 capabilities for the vPC/vPC+ keepalive from the mgmt0 out-of-band interface and also from the built-in front ports using SVI.
Isolate and Maintenance Mode Enhancement
Provides the ability to gracefully eject a switch and isolate it from the network so that debugging or an upgrade can be performed. The switch is removed from the regular switching path and put into a maintenance mode. Once maintenance on the switch is complete, you can bring the switch into full operational mode.
In service software updates (ISSUs) are limited to the three previous releases.
Improves efficiency in the usage of Multicast Expansion Table (MET) entries in the hardware.
Open Management Infrastructure
Open Management Infrastructure (OMI) is no longer supported.
Password Length Enhancement
The following commands have been added to provide the ability to configure the minimum and maximum length of a password:
userpassphrase min-length length
userpassphrase max-length length
show userpassphrase length
Syslog Message as SNMP Trap
The following features have been added:
User Interface for Persistent Logging
Syslog SNMP Traps
Syslog Message Format
Unified Fabric Solution (previously called Dynamic Fabric Automation (DFA))
This software release is the second release to support enhancements to Cisco's Unified Fabric Solution.
Unified Fabric focuses on simplifying, optimizing, and automating data center fabric environments by offering an architecture based on four major pillars: Fabric Management, Workload Automation, Optimized Networking, and Virtual Fabrics.
Each of these pillars provides a set of modular functions which can be used together, or independently, for ease of adoption of new technologies in the data center environment.
Dynamic Fibre Channel over Ethernet (FCoE) over DFA enables I/O consolidation. It permits both LAN and SAN traffic to coexist on the same switch and the same wire.
FEX Based ACL Classification
The FEX-based ACL Classification feature uses TCAM resources on a FEX to perform ACL-based packet classification of incoming packets on the switch. When QoS policies are processed on a FEX, the policies are enforced on the switch and on the associated FEX or FEXes.
New Hardware Features in Cisco NX-OS Release 7.0(3)N1(1)
Cisco NX-OS Release 7.0(3)N1(1) supports the following new hardware:
Cisco Nexus 2348UPQ FEX (N2K-C2348UPQ)
New Software Features in Cisco NX-OS Release 7.0(2)N1(1)
Dynamic Fabric Automation
This software release is the first release to support Cisco's Evolutionary Data Center Fabric solution called Dynamic Fabric Automation (DFA). DFA is evolutionary and is based on the industry leading Unified Fabric solution.
DFA focuses on simplifying, optimizing and automating data center fabric environments by offering an architecture based on four major pillars namely Fabric Management, Workload Automation, Optimized Networking and Virtual Fabrics. Each of these pillars provide a set of modular functions which can be used together or independently for easiness of adoption of new technologies in the data center environment.
The ACL logging feature allows you to monitor IPv6 ACL flows and to log dropped packets on an interface.
Dynamic FCoE Using FabricPath
Dynamic FCoE extends the capability and reliability of storage networks by leveraging FabricPath technology to create logical separation of SAN A and SAN B. FCoE VFCs and Interswitch-Links (ISLs) are dynamically configured, simplifying multihop FCoE deployments in leaf-spine topologies.
New Software Features in Cisco NX-OS Release 7.0(0)N1(1)
Cisco NX-OS Release 7.0(0)N1(1) is a major release that includes bug fixes and the following software features and enhancements:
Allows you to add more nodes at the spine layer as the numbers of servers increases.
Early Warning for Forwarding Information Base Exhaustion
When the Forwarding Information Base (FIB) table is 100% full, the following messages is displayed:
FIB_TCAM_RESOURCE_EXHAUSTION:FIB TCAM usage is at 90 percent.
Explicit Congestion Notification with Weighted Random Early Detection
Explicit Congestion Notification (ECN) with Weighted Random Early Detection (WRED) solves the delay and packet loss problems for applications that are sensitive to these issues.
FabricPatch Operations, Administration, and Management
Support for Fabric Path Operations, Administration and Management has been added in this software release.
Fibre Channel and Fibre Channel Over Ethernet Slow Drain
Fiber Channel (FC) and Fibre Channel over Ethernet (FCoE) slow drain addressed the issue of slow drain devices that cause congestion in the network.
A Multi-Destination Tree (MDT), also referred to as a forwarding tag or ftag, is a spanning-tree used for forwarding packets within a topology. By default, a topology has two MDTs/ ftags: topology 0 has ftag 1 and 2, topology 1 has ftag 3 and 4, topology 2 has ftag 5 and 6, up to a maximum supported 64 topologies.
The OpenFlow feature is a specification from the Open Networking Foundation (ONF) that defines a flow-based forwarding infrastructure (L2-L4 Ethernet switch model) and a standardized application programmatic interface (protocol definition) to learn capabilities, add and remove flow control entries and request statistics. OpenFlow allows a controller to direct the forwarding functions of a switch through a secure channel.
One Platform Kit (OnePK)
Support has been added for One Platform Kit (onePK) Turbo API. OnePK is a cross-platform API and software development kit that enables you to develop applications that interact directly with Cisco networking devices. onePK provides you access to networking services by using a set of controlled APIs that share the same programming model and style. For more information, see the following URL:
Intermediate System to Intermediate System (IS-IS) uses the overload bit to tell other routers not to use the local router to forward traffic but to continue routing traffic destined for that local router.
Port Channel Max Links
The Port Channel Max Links feature defines the maximum number of bundled ports allowed in an LACP port channel.
Switch Port Analyzer with Access Control List Filtering
The Switch Port Analyzer (SPAN) with Access Control List (ACL) filtering feature allows you to filter SPAN traffic so that you can reduce bandwidth congestion. To configure SPAN with ACL filtering, you use ACL’s for the session to filter out traffic that you do not want to span. An ACL is a list of permissions associated to any entity in the system; in the context of a monitoring session, an ACL is a list of rules which results in spanning only the traffic that matches the ACL criteria, saving bandwidth for more meaningful data. The filter can apply to all sources in the session.
You can create and administer up to 16 templates to resize the regions in ternary content-addressable memory (TCAM).
Upgrading or Downgrading to a New Release
This section describes the upgrade and downgrade paths that are supported for Cisco NX-OS Release 7.1(0)N1(1a) on the Cisco Nexus device.
The following guidelines apply to Cisco NX-OS Release 7.1(0)N1(1b) for Cisco Nexus devices:
When a Layer 3 license is installed, the Cisco Nexus 5500 Platform does not support an ISSU. Hot swapping a Layer 3 module, for example, the Layer 3 GEM (N55-M160L3-V2) or Version 2 Layer 3 daughter card (N55-D160L3-V2), is not supported. You must power down the Cisco Nexus device before removing or inserting a Layer 3 expansion module.
Supported Upgrade and Downgrade Paths
Table 4 shows the upgrade and downgrade possibilities for Cisco NX-OS Release 7.1(0)N1(1b). For more information, see the Cisco Nexus 5500 Series NX-OS Software Upgrade and Downgrade Guide, Release 7.1(0)N1(1b).
19.If there are unified ports configured as fiber channel (FC) and a disruptive upgrade is performed, then the FC interfaces must be reconfigured, and the switch will require a second reload.
This section describes the limitations for Cisco NX-OS Release 7.1(0)N1(1b).
Loading a new license or reloading existing license on a Cisco Nexus 5624Q switch is not supported. For details, see CSCus41273.
Ingress inter-VLAN-routed Layer3 multicast packets are treated as “unknown multicast” by the storm-control feature. This is due to the Layer 3 forwarding design in the Cisco Nexus 5500 Series switch. For details, see CSCuh34068.
The Server Virtualization Switch (SVS) connection is not deleted during a rollback when NIV is enabled. To resolve this issue, delete the current SVS connection and reapply the original SVS connection. For details, see CSCts17033.
If SPAN traffic is rate-limited by entering the switchport monitor rate-limit 1G command, then a maximum transmission unit (MTU) truncation size cannot be used to truncate SPAN packets. For details, see CSCua05799.
When an FC SPAN destination port is changed from SD to F mode and back to SD mode on an NPV switch, the port goes into an error-disabled state. Perform a shut/no-shut after the mode change recovers the port. This issue occurs only in NPV mode. For details, see CSCtf87701.
If you configure a Cisco Nexus 2248TP port to 100 Mbps instead of autonegotiation, then autonegotiation does not occur, which is the expected behavior. Both sides of the link should be configured to both hardwired speed or both autonegotiate.
no speed —Autonegotiates and advertises all speeds (only full duplex).
speed 1000 —Autonegotiates only for an 802.3x pause.
speed 100 —Does not autonegotiate; pause cannot be advertised. The peer must be set to not autonegotiate and to fix at 100 Mbps (similar to the N2248TP)
For details, see CSCte81998.
Given the implementation of a single CPU ISSU, the STP root on the PVST region with switches on an MST region is not supported. The PVST simulation on the boundary ports goes into a PVST SIM inconsistent blocked state that breaks the STP active path. To work around this issue, move all STP roots to the MST region. However, the workaround causes a nondisruptive ISSU to fail because nonedge designated forwarding ports are not allowed for an ISSU. For additional information, see CSCtf51577.
IGMP queries sent in CSCtf94558 are group-specific queries that are sent with the destination IP/MAC address as the group's address.
GS queries are sent for IP address: 18.104.22.168 to 22.214.171.124 [0100.5E01.0E01 to 0100.5E01.0E64]
These are not link-local addresses. By default, they are not flooded by the hardware into the VLAN. They are sent only to the ports that have joined this group.
This is expected behavior during an ISSU.
In another scenario, the IGMP global queries [dest IP 126.96.36.199] get flooded correctly in the VLAN.
Group-specific queries are not forwarded to ports other than the one that joined the group during ISSU. The reason to forward group-specific queries toward hosts is to avoid having them leave the group. However, if a port has not joined the group, then this is not an issue. If there is an interface that has joined the group, the queries are expected to make it to the host. While the behavior is different when ISSU is not occurring, it is sufficient and works as expected and there is no impact to the traffic. For details, see CSCtf94558.
When a private VLAN port is configured as a TX (egress) SPAN source, the traffic seen at the SPAN destination port is marked with the VLAN of the ingressed frame. There is no workaround.
In large-scale configurations, some Cisco Nexus 2000 Series Fabric Extenders might take up to 3 minutes to appear online after entering the reload command. A configuration can be termed large-scale when the maximum permissible Cisco Nexus 2000 Series Fabric Extenders are connected to a Cisco Nexus device, all host-facing ports are connected, and each host-facing interface has a large configuration (that supports the maximum permissible ACEs per interface).
Egress scheduling is not supported across the drop/no-drop class. Each Fabric Extender host port does not support simultaneous drop and no drop traffic. Each Fabric Extender host port can support drop or no drop traffic.
The Cisco Nexus 2148 Fabric Extender does not support frames with the dot1q vlan 0 tag.
VACLs of more than one type on a single VLAN are unsupported. Cisco NX-OS software supports only a single type of VACL (either MAC, IPv4, or IPv6) applied on a VLAN. When a VACL is applied to a VLAN, it replaces the existing VACL if the new VACL is a different type. For instance, if a MAC VACL is configured on a VLAN and then an IPv6 VACL is configured on the same VLAN, the IPv6 VACL is applied and the MAC VACL is removed.
A MAC ACL is applied only on non-IP packets. Even if there is a match eth type = ipv4 statement in the MAC ACL, it does not match an IP packet. To avoid this situation, use IP ACLs to apply access control to the IP traffic instead of using a MAC ACL that matches the EtherType to IPv4 or IPv6.
Multiple boot kickstart statements in the configuration are not supported.
If you remove an expansion module with Fibre Channel ports, and the cable is still attached, the following FCP_ERRFCP_PORT errors appear:
These messages are informational only and result in no loss of functionality.
If you configure Multiple Spanning Tree (MST) on a Cisco Nexus device, we recommend that you avoid partitioning the network into a large number of regions.
By design, vEth interfaces do not share the underlying behavior of a vPC port. As a result, a VLAN does not get suspended when the peer switch suspends it. For example, when you shut a VLAN on a primary switch, the VLAN continues to be up on the secondary switch when the vEth interface is on a FEX. When the VLAN on the primary switch goes down, the VLAN on the vEth interface on the primary is suspended, but the vEth on the secondary switch is up because it is an active VLAN on the secondary switch.
Role-based Access Control List (RBACL) policy enforcement is performed on VLANs on which Cisco Trusted Security (CTS) enforcement is not configured. This situation occurs when there is at least one VLAN in the switch where CTS is enforced. On a VLAN where CTS is not enforced, RBACL policy lookup occurs for ingress packets and the packet is denied or permitted according to the policies in the system. To work around this issue, make sure that all VLANs on which SGT tagged packets ingress enforce CTS.
The packet length in the IP GRE header of a packet exiting from the switch is not equal to the MTU value configured in the ERSPAN source session. This is true for SPAN or ERSPAN. This situation can occur whenever the MTU value that is configured in an ERSPAN or SPAN session is smaller than the SPAN packet, such as when the packet is truncated. The IP GRE packet is truncated to a value that differs by –2 to 10 bytes from the expected MTU.
When you configure a Layer 3 interface as an ERSPAN source, and configure the ERSPAN termination on a Catalyst 5500 switch or a Cisco Nexus 7000 Series switch, you cannot terminate the Layer 3 interface ERSPAN source on the Cisco Nexus 7000 Series switch or the Catalyst 5500 switch. To work around this issue, configure VLAN 1 to 512 on the Cisco Nexus 7000 Series switch or the Catalyst 6000 switch.
Unknown unicast packets in FabricPath ports are counted as multicast packets in interface counters. This issue occurs when unknown unicast packets are sent and received with a reserved multicast address (that floods to a VLAN) in the outer FabricPath header, and the Cisco Nexus device increments the interface counter based on the outer FabricPath header. As a result, multicast counters are incremented. In the case of a Cisco Nexus 7000 Series switch, unicast counters are incremented as they are based on an inner Ethernet header. There is no workaround for this issue.
In an emulated switch setup, inband keepalive does not work. The following steps are recommended for peer keepalive over switch virtual interface (SVI) when a switch is in FabricPath mode:
– Use a dedicated front panel port as a vPC+ keepalive. The port should be in CE mode.
– Use a dedicated VLAN to carry the keepalive interface. The VLAN should be a CE VLAN.
– Add the management keyword to the corresponding SVI so that the failure of a Layer 3 module will not bring down the SVI interface.
– Enter the dual-active exclude interface-vlan keepalive-vlan command to prevent the SVI from going down on the secondary when a peer-link goes down.
FabricPath requires 802.1Q tagging of the inner Ethernet header of the packet. Native VLAN packets that are sent by a Cisco Nexus 7000 Series switch are not tagged. As a result, a Cisco Nexus device drops packets due to packet parsing errors. To work around this issue, enter the vlan dot1q tag native command on the Cisco Nexus 7000 Series switch to force 802.1Q tagging of native VLAN packets.
A nondisruptive ISSU is not supported when ingress policing is configured.
The maximum IP MTU that can be set on Layer 3 interfaces on which Layer 3 protocols are running is 9196, because of the internal header used inside the switch. The network-qos policy must be set to 9216.
If there are unified ports configured as fiber channel (FC) and a disruptive upgrade is performed, then the FC interfaces must be reconfigured, and the switch will require a second reload.
Limitations on the Cisco Nexus Device
The limitations on the Cisco Nexus device 5500 Series devices are as follows:
The SPAN limitations on Fabric Extender ports are as follows:
On a Cisco Nexus device, if the SPAN source is a FEX port, the frames will always be tagged when leaving the SPAN destination.
On a Cisco Nexus 5500 Platform switch, if the SPAN source is on an access port on the switch port, the frames will not be tagged when leaving the SPAN destination.
Ports on a FEX can be configured as a tx-source in one session only.
If two ports on the same FEX are enabled to be tx-source, the ports need to be in the same session. If you configure a FEX port as a tx-source and another port belonging to the same FEX is already configured as a tx-source on a different SPAN session, an error is displayed on the CLI.
In the following example, Interface Ethernet100/1/1 on a FEX 100 is already configured as a tx-source on SPAN session-1:
swor28(config-monitor)# show running-config monitor
monitor session 1
source interface Ethernet100/1/1 tx
destination interface Ethernet1/37
If you add an interface Ethernet100/1/2 as a tx-source to a different SPAN session (session-2) the following error appears:
ERROR: Eth100/1/2: Ports on a fex can be tx source in one session only
When a FEX port is configured as a tx-source, the multicast traffic on all VLANs for which the tx-source port is a member, is spanned. The FEX port sends out only multicast packets that are not filtered by IGMP snooping. For example, if FEX ports 100/1/1–12 are configured on VLAN 11 and the switch port 1/5 sends multicast traffic on VLAN 11 in a multicast group, and hosts connected to FEX ports 100/1/3-12 are interested in receiving that multicast traffic (through IGMP), that multicast traffic goes out on FEX ports 100/1/3–12, but not on 100/1/1–2.
If you configure SPAN Tx on port 100/1/1, although the multicast traffic does not egress out of port 100/1/1, the SPAN destination does receive that multicast traffic, which is due to a design limitation.
When a FEX port is configured as both SPAN rx-source and tx-source, the broadcast, non-IGMP Layer-2 multicast, and unknown unicast frames originating from that port might be seen twice on the SPAN destination: once on the ingress and once on the egress path. On the egress path, the frames are filtered by the FEX to prevent them from going out on the same port on which they were received. For example, if FEX port 100/1/1 is configured on VLAN 11 and is also configured as SPAN rx-source and tx-source and a broadcast frame is received on that port, the SPAN destination recognizes two copies of the frame, even though the frame is not sent back on port 100/1/1.
A FEX port cannot be configured as a SPAN destination. Only a switch port can be configured and used as a SPAN destination.
Layer 3 Limitations
In a vPC topology, two Cisco Nexus devices configured as vPC peer switches need to be configured symmetrically for Layer 3 configurations such as SVIs, the peer gateway, routing protocol and policies, VRF, and RACLs.
Note The vPC consistency check does not include Layer 3 parameters.
When a Layer 3 module goes offline, all non-management SVIs are shut down. To maintain connectivity when a Layer 3 module fails, you can configure an SVI as a management SVI using the command management under i nterface vlan. This prevents traffic to the management SVI from passing through the failed Layer 3 module.
Cisco Nexus 5548P Daughter Card (N55-D160L3)
Before installing a Layer 3 daughter card (N55-D160L3) into a Cisco Nexus 5548P switch, you must upgrade to Cisco NX-OS Release NX-OS Release 5.0(3)N1(1c) or a later release, and then install the card into the chassis.
This section includes the open and resolved caveat record numbers for this release. Links are provided to the Bug Toolkit where you can find details about each caveat.
dot1x: traffic flooding due to miss mac address in MAC table
The Cisco Management Information Base (MIB) list includes Cisco proprietary MIBs and many other Internet Engineering Task Force (IETF) standard MIBs. These standard MIBs are defined in Requests for Comments (RFCs). To find specific MIB information, you must examine the Cisco proprietary MIB structure and related IETF-standard MIBs supported by the Cisco Nexus 5500 Series switch.
The MIB Support List is available at the following FTP site:
The documentation set includes the following types of documents:
Licensing Information Guide
Installation and Upgrade Guides
Configuration Examples and TechNotes
Error and System Message Guides
Security Advisories, Responses and Notices
MIB Reference Guide
To provide technical feedback on this document or to report an error or ommission, please send your comments to firstname.lastname@example.org. We appreciate your feedback.
Cisco and the Cisco logo are trademarks or registered trademarks of Cisco and/or its affiliates in the U.S. and other countries. To view a list of Cisco trademarks, go to this URL: www.cisco.com/go/trademarks. Third-party trademarks mentioned are the property of their respective owners. The use of the word partner does not imply a partnership relationship between Cisco and any other company. (1110R)