Configuration Guide User guide

850 FastIron Configuration Guide
53-1002494-02
Layer 2 behavior with MCT
22
MCT Layer 2 protocols
Keep the following information in mind when configuring Layer 2 protocols with MCT.
MRP
• An ICL interface cannot be configured as an MRP secondary interface or vice versa, because
the ICL cannot be BLOCKING.
• MRP cannot be enabled on MCT CCEP port and vice versa.
STP/RSTP
• STP is not recommended to be configured on MCT VLANs at MCT cluster devices. By default,
the spanning tree is disabled in the MCT VLANs. If the network topology may be creating Layer
2 loops through external connections, STP could be enabled on switches outside the MCT
cluster to prevent the Layer 2 loop. The MCT cluster devices will perform a pass-through
forwarding of STP BPDUs received through its ports in the MCT VLAN.
• In rare cases in which the network topology consists of Layer 2 loops outside the MCT cluster
that require STP/RSTP to be enabled on MCT VLANs in the cluster, the CCEPs will always be in
the spanning tree disabled state.
• The STP/RSTP algorithms have been modified such that ICL never goes to blocking. The ICL
guard mechanism ensures that if ICL is going into a blocking state, then the port on which the
superior BPDUs are being received is moved to blocking state and the ICL guard timer starts
running on it. This timer runs as long as superior BPDUs are received on this interface. As long
as this timer runs on an interface, the superior BPDUs are dropped.
• The new BLK_BY_ICL STP state indicates that the superior BPDUs were received on this
interface, which could have led to blocking of the ICL interface, with the result that the CL port
guard mechanism has been triggered on this port.
• In a 802.1s MSTP deployment, Brocade recommends disabling spanning tree on the MCT
cluster devices at the global level. MSTP cannot be configured on individual cluster devices.
• An MCT cluster can support up to 32 spanning tree instances.
BPDU forwarding
If the network deploys single STP or IEEE 802.1s (MSTP), both the MCT cluster devices must be
configured using the bpdu-flood-enable command to flood the single STP/MSTP BPDUs in the
SSTP/MSTP domain (forward to all of the ports in the cluster switch irrespective of VLAN.)
Syntax: [no] bpdu-flood-enable
When bpdu-flood-enable is configured, there should not be any links other than the ICL (including
the keep-alive VLAN link) connecting the two MCT cluster devices. If there is an additional link,
then the flooded BPDU will cause a loop and high CPU utilization.
Protocol-based VLANs
Protocol and subnet VLANs can be configured on MCT VLANS, however, ICL and CCEPs cannot be
configured as dynamic members of protocol based VLANs (and vice versa). ICL and CCEP can
either be excluded or static members of protocol based VLANs. CEPs can be configured as dynamic
or static, or exclude members of protocol based VLANs.