This KB article describes some of the MC-lag failure handling scenarios.

MC-LAG TOPOLOGY:

SWITCH 1 <---ae0 (iccp/icl)---> SWITCH2

CONFIGURATION:

ICCP configuration on the box: (similar config for SWITCH-2)

root@SWITCH-1> show configuration protocols iccp

local-ip-addr 192.168.1.X;

peer 192.168.1.Y {

   redundancy-group-id-list 201;

   backup-liveness-detection {

       backup-peer-ip 1.1.1.Y;

   }

   liveness-detection {

       minimum-receive-interval 1000;

       transmit-interval {

           minimum-interval 1000;

       }

   }

}

• ae0 is used for both iccp and icl .

root@SWITCH-1# run show configuration | display set| match ae0

set interfaces xe-0/0/X ether-options 802.3ad ae0

set interfaces xe-0/0/X ether-options 802.3ad ae0

set interfaces ae0 aggregated-ether-options lacp active

set interfaces ae0 unit 0 family ethernet-switching interface-mode trunk

set interfaces ae0 unit 0 family ethernet-switching vlan members all

set multi-chassis multi-chassis-protection 192.168.1.Y interface ae0

• Backup-liveness-detection IP is through em0

LAB TESTING

Scenario 1: 

• Disable all interfaces on Switch 1 : 

• As the Backup-liveness status is up . The device considers this as a split-brain state 
• LACP system ID is changed on both Active and stand-by .
• This will be expected behavior . 

Scenario 2 

• Disable all interfaces on Switch 1 and the Backup-liveness Status is down .
• This means entire switch Sw01is down. 

Scenario 3:

• Disable all interfaces on Sw2 .

• As the Backup-liveness status is up . The device considers this as a split-brain state 
• LACP system ID is changed on both Active and stand-by .
• This will be expected behavior . 

Scenario 4 : 

• Disable all interfaces on Sw02 . 
• The below commands are configured on Sw1 : 

root@SWITCH-01> show configuration | display set | match peer  

set interfaces ae11 aggregated-ether-options mc-ae events iccp-peer-down prefer-status-control-active

set interfaces ae12 aggregated-ether-options mc-ae events iccp-peer-down prefer-status-control-active

set interfaces ae21 aggregated-ether-options mc-ae events iccp-peer-down prefer-status-control-active

set interfaces ae22 aggregated-ether-options mc-ae events iccp-peer-down prefer-status-control-active

• Though the device considers this as a split-brain state but since we had, “prefer-status-control-active” knob enabled, the LACP system ID does not change. 

Scenario 5:

• Disable all interfaces on Switch 2 and the Backup-liveness Status is down .
• This means entire switch Sw02 is down. 

 Refer DOC : https://www.juniper.net/documentation/us/en/software/junos/mc-lag/topics/concept/best-practices-usage-notes.html