Back to Back MC-LAG without static ARPs might cause an ARP storm.

DDOS ARP is seeing causing network issues.

Configure static ARP on the ICCP peers.

Additional configuration has been suggested along with static ARP for ICCP.

Step 1: Add bfd interval to 8 seconds on both mc ae nodes :

MC AE node 1

root@MC-1# show configuration protocols iccp

local-ip-addr 10.10.10.1;

  peer 10.10.10.2 {

   session-establishment-hold-time 340;

   redundancy-group-id-list 1;

     liveness-detection {

     minimum-interval 2000;

          multiplier 4;

    }

  }

root@MC-1# commit

MC-AE node 2

root@MC-2# show configuration protocols iccp

local-ip-addr 10.10.10.2;

 peer 10.10.10.1 {

  session-establishment-hold-time 340;

  redundancy-group-id-list 1;

  liveness-detection {

      minimum-interval 2000;

          multiplier 4;

   }

 }

• Commit

Step 2 :Removing “arp-l2-validate” on both mc-ae nodes

Node 1

 delete interface irb arp-l2-validate

 commit

Node 2

delete interface irb arp-l2-validate

 commit

Step 3: Adding static ARP for ICCP on both nodes

     Node 1

root@MC-1# show interfaces irb.100

 family inet {

  address <10.10.10.1/30> {

    arp <10.10.10.2> l2-interface ae100.0 mac <peer node IRB mac>;

}

root@MC-1# commit

Node 2

root@MC-2# show interfaces irb.100

 family inet {

  address <10.10.10.2/30> {

    arp <10.10.10.2> l2-interface ae100.0 mac <peer node IRB mac>;

root@MC-2# commit

    Step 4 : Adding prefer-status-control active for ICCP down protection (on node with status-control-active)

• For every mc-ae with status-control-active , add prefer-status-control-active as below :

 For example in one ae in MC-1

• For mc-ae with static-control-active add the below

master:0}[edit interfaces ae1]

root@MC-1# show

aggregated-ether-options {

  lacp {

      active;

      system-id 01:00:00:00:00:01;

      admin-key 1;

  }

  mc-ae {

      mc-ae-id 1;

      redundancy-group 1;

      chassis-id 0;

      mode active-active;

      status-control active;

      init-delay-time 240;

  }

}

unit 0 {

  family ethernet-switching {

      interface-mode trunk;

      vlan {

          members all;

      }

  }

}

{master:0}[edit interfaces ae1]

root@MC-1# set aggregated-ether-options mc-ae events iccp-peer-down prefer-status-control-active

root@MC-1# commit 

https://www.juniper.net/documentation/us/en/software/junos/mc-lag/topics/topic-map/examples-mc-lag.html