Whenever an edge/designated port changes status, either due to up/down or admin disable/enable, the root port (uplink to the root bridge) enters discarding mode for MSTI 0 (and any other configured instances). Then quickly transitions back into learning, then forwarding.
This results in interruptions on any frames traversing the switch.
Switching Infrastructure is as follows;
Root Bridge;
1 x Dell S4128F-ON (Priority 4096 // 0x1000)
1 x Dell S4128F-ON (Priority 8192 // 0x2000)
Member Bridges;
1 x HP Procurve 2848 (Priority 40960 / 0xA000)
4 x MikroTik CRS354 (Priority 28672 / 0x7000)
oct/29 20:14:06 interface,info ether17 link down
oct/29 20:14:06 bridge,stp sfp-sfpplus2:0 discarding
oct/29 20:14:06 bridge,stp sfp-sfpplus2:1 discarding
oct/29 20:14:08 bridge,stp sfp-sfpplus2:0 learning
oct/29 20:14:08 bridge,stp sfp-sfpplus2:0 forwarding
oct/29 20:14:08 bridge,stp sfp-sfpplus2:1 learning
oct/29 20:14:08 bridge,stp sfp-sfpplus2:1 forwarding
..
Code: Select all
[admin@xxx-xxx-sw01] > interface bridge monitor br0
state: enabled
current-mac-address: C4:AD:34:C4:DE:BA
root-bridge: no
root-bridge-id: 0x1000.68:4F:64:57:50:15
regional-root-bridge-id: 0x1000.68:4F:64:57:50:15
root-path-cost: 0
root-port: sfp-sfpplus2
port-count: 60
designated-port-count: 20
mst-config-digest: c4073f30b644e468297a52c7af83ee4a
fast-forward: no
Code: Select all
[admin@xxx-xxx-sw01] > interface bridge msti monitor 0
state: enabled
identifier: 1
current-mac-address: C4:AD:34:C4:DE:BA
root-bridge: no
root-bridge-id: 0.00:00:00:00:00:00
regional-root-bridge-id: 0x1001.68:4F:64:57:50:15
root-path-cost: 0
root-port: sfp-sfpplus2
port-count: 60
designated-port-count: 20
Configuration is as follows
Code: Select all
/interface bridge
add auto-mac=no comment=defconf name=br0 priority=0x7000 protocol-mode=mstp region-name=Xxxx vlan-filtering=yes
/interface bridge msti
add bridge=br0 identifier=1 priority=0x7000 vlan-mapping=1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16
The following Firmware has been tested.
6.47.4
6.46.1
7.1 Beta 2 (wanted to check on a newer kernel)
We have replicated this behaviour on a single switch.
This behaviour is also exhibited regardless of whether the port is up or down and is disabled / enabled.
This seems to happen regardless of Edge port configuration, BPDU guard on 'designated' ports.
This is not expected behaviour.
Is this something to do with the switch chipset or the bridge implementation?