Description

This article will discuss the behavior of VRRP when different virtual IPs are configured within the same group.

Symptoms

We have captured the errors on Switch1-DC2 (Please see the topology below).

 

##########################

 

Sep 14 19:30:46.812 Switch1-DC2 vrrpd[29426]: VRRPD_MISSING_VIP: Packet received by irb.2003 for VRRP group 1 did not include required virtual IP address 2123:1234:1234:600::1

Sep 14 19:30:47.094 Switch1-DC2 vrrpd[29426]: VRRPD_MISSING_VIP: Packet received by irb.2004 for VRRP group 1 did not include required virtual IP address 2123:1234:1234:601::1

 

 

Switch1-DC2> show vrrp extensive 

....... irb.2003 ......

  Group VRRP PDU statistics
    Advertisement sent                       :300122
    Advertisement received                   :300607
  Group VRRP PDU error statistics
    Bad authentication Type received         :0
    Bad password received                    :0
    Bad MD5 digest received                  :0
    Bad advertisement timer received         :0
    Bad VIP count received                   :0
    Bad VIPADDR received                     :300604 <<<<<<<<<<
  Group state transition statistics
    Idle to master transitions               :0
    Idle to backup transitions               :1
    Backup to master transitions             :1
    Master to backup transitions             :0

 

 

Its happening on both Switch1-DC2 and Switch2-DC2

 

 

Topology:-

 

 

 

 

We have four switches within the same broadcast domain (VLAN 2003): Switch1-DC2 and Switch2-DC2 in Data Center 2, and Switch1-DC1 and Switch2-DC1 in Data Center 1. Although these are different PODs, VLAN 2003 spans across all four switches. We are using the same group name ID and the same IRB (.2003) for all devices.

 

 

Switch1-DC1

 irb {

    unit 2003 {

      family inet6 {

        address 2123:1234:1234:100::2/64 {

          vrrp-inet6-group 1 { <<<<<<<<<<< Same group id

            virtual-inet6-address 2123:1234:1234:100::1; <<<<<<< Different subnet ip address as Per Switch1-DC2

 

Switch2-DC1

 irb {

    unit 2003 {

      family inet6 {

        address 2123:1234:1234:100::3/64 {

          vrrp-inet6-group 1 { <<<<<<<<<<< Same group id

            virtual-inet6-address 2123:1234:1234:100::1; <<<<<<< Different subnet ip address as Per Switch1-DC2

 

            

Switch1-DC2

 

irb {

    unit 2003 {

      family inet6 {

        address 2123:1234:1234:600::3/64 {

          vrrp-inet6-group 1 { <<<<<<<<<<< Same group id

            virtual-inet6-address 2123:1234:1234:600::1;

 

Switch2-DC2

 

 irb {

    unit 2003 {

      family inet6 {

        address 2123:1234:1234:600::2/64 {

          vrrp-inet6-group 1 { <<<<<<<<<<< Same group id

            virtual-inet6-address 2123:1234:1234:600::1;

     

When the VRRP PDUs are generated by the VRRP master in DC1, they are flooded to both Switch1 and Switch2 in DC2. Similarly, when the VRRP master in DC2 generates PDUs, they are flooded to Switch1 and Switch2 in DC1. We have confirmed this through tcpdump/pcap captures taken on IRB .2003 on the DC2 switches, and the VRRP PDU statistics further support this observation.

 

 

 

In this case, we have two different VRRP virtual IPs configured within the same VRRP group on the same VLAN. This explains the presence of the VRRPD-MISSING-VIP messages on all four devices.

 

Even though the Layer 3 subnets are configured differently in DC1 and DC2, the VRRP PDUs are processed at Layer 2 and are flooded across the same domain.

 

 

 

Solution

The recommendation is to change the VRRP group ID on either DC1 or DC2 to a different value.

 

Alternatively, we can configure the same VIP within the same IRB.

 

Another option is to stop communication between IRB .2003 and IRB .2004 between DC1 and DC2.

Modification History

2024-09-26 : Article Created