Previously, cluster STP operation does not support fast failover because the STP state machine state is not shared between the two nodes. As a result, when the master fails, the slave recomputes the STP state from scratch. When the cluster comes online, the cluster recomputes the STP state from scratch. This causes topology changes which cause traffic loss until the STP converges. Fast failure is now supported by default.
New show commands are available for this feature:
CLI network-admin@Leaf1 > stp-state-show
STP parameters such as bridge-priority, port cost values configured before upgrading to Netvisor OS 2.4.0 are set to default values after upgrade to Netvisor OS 2.4.0. You must reconfigure STP after upgrading the software.
Multiple Spanning Tree Protocol as defined in IEEE802.1s or IEEE802.1Q-2005 provides the ability to manage multiple VLANs from a single Multiple Spanning Tree (MST) instance. MST allows the formation of MST regions that can run multiple MST instances (MSTIs). Multiple regions and other STP bridges are interconnected using one single common spanning tree (CST).
MSTP regions are be defined to be a collection of switches that must have the same VLANs on all of the switches. Each MST region must a root bridge. The root bridge may not reside outside of the region. MST for a single region is supported. This enables multiple MST instances within a single region.
The following commands support the configuration of MST instances on a local switch:
CLI network-admin@Leaf1 > mst-config-create
Specify the ID as a number between 0 and 63 for MST configuration.
Specify the list of VLANs associated with the MST configuration
Specify the bridge priority number for MST.
Additional commands for MST include the following:
To configure MST, use the following commands:
CLI network-admin@Leaf1 > mst-config-create instance-id 1-63 vlans vlan-list bridge-priority 4096
The bridge priority is a value from 0 to 65536, with a default value of 0. The value increments by 4096 each time. For example, the values can be 0, 4096, 8192, up to 65536.
Port hairpinning allows Layer 2 bridged traffic to exit out of the same switch-port that it arrived on. This is useful because it supports hosting containers with Single Root I/O Virtualization (SR-IOV) network interfaces and classifies traffic going towards the applications.
This feature also allows the first-hop switch to enforce policies and security rules in hardware, through vflows, and may be used where a Netvisor OS-enabled switch is used to micro-segment traffic, such as whitelists
You can use this feature when modifying a port configuration and when creating or modifying a trunk configuration with link aggregation.
Informational Note: If you configure this feature on a port that is not connected to a server, it may cause network issues.
The following types of traffic to bridge back:
Layer 2 Unicast traffic
Layer 2 Broadcast, Unknown Unicast, Multicast (BUM) traffic
CPU originated packets
To enable this feature, use the following command:
CLI network-admin@Leaf1 > port-config-modify port port-list reflect
To disable this feature:
CLI network-admin@Leaf1 > port-config-modify port port-list no-reflect