Best practices /Troubleshooting MST/PVST interaction

Recommendation/ Best Practices/Design recommendation.


    1.    Use Rapid PVST instead PVST on all the switches to make spaning-tree convergence faster.
    2.    Make Core switch as spaning-tree  root bridge by making highest spanning-tree priority on
           switch  since core switch connects to all distribution.
    3.    To prevent other switch to become spanning-tree root bridge, always  apply spanning-tree  
           root  guard on all downstream links connecting to distribution switch.
    4.    Enable spanning-tree portfast bpduguard  on all the  port on access switches.
    5.    Etherchannel misconfig guard has already been enabled.
    6.    Root bridge  for all VLANS should be in MST region when interacting with PVST+/RPVST+
           boundary.



Commands to use while troubleshooting.

Show spanning-tree mst details.
Clear spanning-tree counters
show spanning-tree mst configuration.
Show spanning-tree details | i BPDU|from|exec|ieee|last



Facts :

1) MST runs MST0 (one instance for all  VLANS even when there is no VLAN mapped to MST0 instance) for interacting with PVST+/RPVST switches.
2) PVST+/RPVST+ switches always follow the priority of MST0 instance and its properties.
3) MST0 (Commom spanning tree or CIST root bridge) informed other MST 1 or another MST instance about the TCN changes received from PVST+/RPVST+ switches.
4) PVST+/RPVST+ switches sees entire MST region as one PVST /RPSVT switch bridge.
5) If MST instance is not configured correctly , it sees as RSTP boundary ports.
6) During testing, it is found that native vlan not used for MST BPDU- still  have doubt on it.
7) In PVST , whenever TCN generates  by switch, it informs the neigbor switch towards the ROOT bridge and neigbor switch ackwnowledge the TCN BPDU to switch to make him stop the BPDU. Same process goes up to the root bridge and then root bridge starts sending the TCN BPDU to inform all the switch to reduce the aging timer to forward timer and flush the mac address table.
8) In contrast to PVST, RSTP behaved differently in TCN. Whenever TCN is generated in RPVST, it flush the mac address table and inform other switches at the same.

Facts and Troubleshooting IGMP and IGMP snooping

IGMP and IGMP snooping both are different protocol. IGMP is used between host and multicast router for  multicast membership communication  and IGMP snooping is used on switch /vlan to limit forwarding multicast traffic to only interested host on switch.


Behavior on Cat4948/15.0(2)SG5

1) show ip igmp groups ----output for v2 and version 3 only if pim sparse mode is configured under the interface at layer 3.


2) Show ip igmp membership ---output for version 3. version 3 needs to be configured.


3) show ip igmp snooping  groups --------output for v2 and version 3.


4) Show ip gimp snooping membership- No output at all even with multiple images. Explicit tracking must be enabled for output for command 4 .


5)  Scenario -1 :  If igmp snooping is disabled, it will flood the traffic .  Enabling of igmp at layer 3 interface like pim -sparse mode does not affect.

6) Scenario -2 : If igmp snooping is enabled and there is no entry in the snooping table for particular  multicast group, it will not flood the traffic and looks like it is get dropped.


7) show ip gimp snooping mrouter


8) show ip gimp snooping querier.

Note: In source only vlan, traffic is only forwarded to mrouter port if igmp snooping is enabled.

Note : Whenever we configure the igmp version 3, host on the LAN automatically switches to version 3 and start sending version 3 membership report. And when we unconfigure , it switches back to  igmp version 2.

Note: Enable igmp querier on same vlan switch if there is no querier.



Behavior on 6500

Note : On cat6k , show ip igmp snooping  groups does not work and need to use the "show mac address-table address multicast igmp-snooping" since forwarding is based on the layer 2 mac-address.