> Source: [sk115274](https://support.checkpoint.com/results/sk/sk115274)

# sk115274 - Disconnecting all cables from all Bond slaves and reconnecting a cable only to a single Bond slave causes both VRRP cluster members to enter Master/Master state

| Property | Value |
|----------|-------|
| Solution ID | sk115274 |
| Date Created | 2016-12-30 |
| Last Modified | 2021-08-10 |
| Technical Level | Advanced |
| Products | Security Gateway |
| Versions | R81.10 (EOS), R81 (EOS) |
| OS | Gaia |

## Symptoms

- * Disconnecting all cables from all Bond slaves on one VRRP cluster member and reconnecting a cable only to a single Bond slave (on that VRRP cluster member) causes both VRRP cluster members to enter Master/Master state in the scenario described below.

* Topology:

  * VRRP cluster members on Gaia OS.
  * Bond interfaces (each with two slaves) are configured in either Active-Active mode, or 802.3ad mode.
  * Two switches in a stacked configuration that are connected together through a backplane acting as a single entity.
  * Spanning Tree Protocol is enabled on the stacked switch.
  * Each slave in each bond is connected to each switch for redundancy.

    ```
    
    [ switch 1 ]=backplane=[ switch 2 ]
       |    |  _______________|    |
       |    |  |                   |
       |    |__|______________     |
       |       |              |    |
     (eth0) (eth1)         (eth0) (eth1)
        {Bond0}              {Bond0}
    [VRRP member 1]       [VRRP member 2]
    ```

* Behavior:

  * If a cable is disconnected from a single slave interface in the Bond, then nothing happens.
  * If a cable is disconnected from the remaining slave interface in the Bond, then fail-over occurs as expected.
* Issue:

  * If a cable is reconnected only to a single slave interface in the Bond, both VRRP cluster members enter a VRRP Master/Master state due to lost VRRP packets (the VRRP packets are getting lost somewhere between the VRRP cluster members).
* Traffic captures during the issue show:

  * On a member, on which a cable was reconnected to a single slave ("Member_A"), there are multiple VRRP announcement packets with the Source IP address of this member
  * On a peer member (on which the cables were not touched; "Member_B"), there are very few or none of the VRRP announcement packets  
    (When filtering the traffic capture for the Source IP address of "Member_B", there are VRRP announcement packets on both ends - so nothing is being lost in the middle)

## Cause

The Bond interface would come up, but the switch would not pass the packets yet (due to a race condition), even though the Bond reported itself as "Up" on the VRRP cluster member.

## Solution

This solution requires authentication. Please log in to view the full solution.

---

# Agent Instructions

This content is from the Check Point Support Center (https://support.checkpoint.com), the official knowledge base for Check Point cybersecurity products.

## Navigating This Knowledge Base

- **Complete index**: [llms.txt](https://support.checkpoint.com/llms.txt)
- **All SK articles**: [SecureKnowledge Sitemap](https://support.checkpoint.com/sitemaps/secureknowledge-sitemap-index.xml)
- **SK article URL pattern**: `https://support.checkpoint.com/results/sk/{skId}`
- **Markdown responses**: AI bot User-Agents automatically receive `text/markdown` content