The following list identifies restrictions and limitations associated with RSTP and MSTP:
RSTP mode does not support SPBM.
A C-VLAN-level loop across SPBM NNI ports cannot be detected and needs to be resolved at the provisional level.
SPBM NNI ports are not part of the Layer 2 VSN C-VLAN, and BPDUs are not transmitted over the SPBM tunnel. SPBM can only guarantee loop-free topologies consisting of the NNI ports. You should always use Simple Loop Prevention Protocol (SLPP) in an SMLT environment.
Note
Deploy SLPP on C-VLANs to detect loops created by customers in their access networks. However, SLPP is not required on B-VLANs, and it is not supported. The B-VLAN active topology is controlled by IS-IS that has loop mitigation and prevention capabilities built into the protocol.
SPB internally uses spanning tree group (STG) 63 or Multiple Spanning Tree Instance (MSTI) 62. STG 63 or MSTI 62 cannot be used by another VLAN or MSTI. For non-SPB customer networks, if you use STG 63 or MSTI 62 in the configuration, you must delete STG 63 or MSTI 62 before you can configure SPBM.
You must configure SPBM B-VLANs on all devices in the same MSTP region. MSTP requires this configuration to generate the correct digest.
Configure the SPBM B-VLANs to use matching VLAN IDs.
Use NNI ports exclusively to transport traffic for SPB-based services and not be configured as members of any VLANs other than SPB B-VLANs. In releases that do not support nni-mstp, when an SPBM IS-IS interface is created on an NNI port or an MLT, MSTP is automatically disabled for MSTI-62 on the port/MLT. However, MSTP is not automatically disabled on NNI ports for the CIST (default MSTI). In releases that support the boot config flags nni-mstp command, the default behavior of the MSTP NNI ports is that CIST is disabled automatically on the NNI and the NNI ports cannot be members of any VLANs other than B-VLANs. The default boot config flags nni-mstp must be set to false (which is the default). The following example shows the command to disable the MSTP on the NNI ports.
Switch:1(config)#interface gigabitEthernet 1/8 Switch:1(config-if)#no spanning-tree mstp
In releases that do not support nni-mstp boot configuration, you can support the coexistence of non-SPB based services on the NNI ports, by adding NNI ports as members of VLANs, except for B-VLANs. These other VLANs rely on the use of MSTP for Loop prevention. The network operator must carefully consider the implications of keeping MSTP enabled on the NNI ports because any MSTP topology changes detected on the NNI ports impacts all services and causes most dynamically learned information on the UNI side to be flushed and relearned. This includes, but is not limited to, all customer MAC and ARP records. This can also cause all the UNI ports on a BEB to be temporarily put into a spanning-tree blocking state before transitioning to a forwarding state again. The net result is that MSTP topology changes on the NNI ports adversely impact traffic for SPB-based services. Therefore, the NNI ports be used exclusively for SPB traffic.
The following list identifies restrictions and limitations associated with SPBM IS-IS:
The switch does not support IP over IS-IS as defined by RFC 1195. IS-IS protocol is only to facilitate SPBM.
The switch uses level 1 IS-IS. The switch does not support level 2 IS-IS. The CLI command show isis int-l2-contl-pkts is not supported because the IEEE 802.1aq standard currently only defines the use of one hierarchy, Level 1.
The IS-IS standard defines wide (32bit ) metrics and narrow (8 bits) metrics. The switch supports the wide metric.
IS-IS enabled ports cannot be added to an MLT. The switch does not support this configuration.
The switch does not support NNI on SMLT links.
VLACP is generally used when a repeater or switch exists between connected switches to detect when a connection is down even when the link LED is lit. If you configure VLACP on an SPBM link that is also an IST link, during a connection fail over (where the link LED stays lit) the IS-IS hellos time out first (after 27 seconds, using default values) and take down the IS-IS adjacency. IS-IS then calculates the new shortest path and fails over the SPBM traffic. 90 seconds after the connection failure (using default values), VLACP goes down but the IST link was already taken down by IS-IS.
In this scenario, there is no data traffic impact because IS-IS can find another path in the SPBM network before VLACP goes down.
On each SPBM peer, if you configure the SPBM B-VLANs to use different VLAN IDs, for example, VLAN 10 and 20 on one switch, and VLAN 30 and 40 on the second, the system does not generate a trap message to alert of the mismatch because the two switches cannot receive control packets from one another. Configure the SPBM B-VLANs to use matching VLAN IDs.
Do not change the system MTU to less than the default value of 1950 bytes. The system MTU must be 1950 or jumbo because of the header size increase when transmitting packets over the SPBM cloud.
IP multicast over Fabric Connect cannot connect to existing Protocol Independent Multicast (PIM) networks that connect to SPB originated streams or that add PIM network streams into the SPB network. SPB-PIM Gateway (SPB-PIM GW), however, provides multicast interdomain communication between an SPB network and a PIM network. SPB-PIM GW accomplishes this interdomain communication across a special Gateway VLAN. The Gateway VLAN communicates with the PIM network through the PIM protocol messaging and translates the PIM network requirements into SPB language, and vice versa. For more information about SPB-PIM GW, see SPB-PIM Gateway configuration.
The following list identifies other restrictions or considerations:
You cannot use 3.33.33 as the SPB nickname because of a conflict with reserved IPv6 Ethernet multicast address 33:33:xx:xx:xx:xx.
The software does not support I-SID filters.
You cannot enable C-VLAN and B-VLAN on the same port.
To ensure proper cleanup of MAC tables after you run the no spbm command, save the configuration, and then reboot the switch.