RabbitMQ Tile Upgrade from 2.3.5 to 2.4.2 Stuck Due to Khepri DB Errors

Products

• Services Suite
• VMware Tanzu RabbitMQ
• RabbitMQ
• VMware Tanzu Data Intelligence
• VMware Tanzu Data Suite
• VMware Tanzu Data Suite

Issue/Introduction

During the upgrade from 2.3.5 to 2.4.2, the following error message appears repeatedly in the RabbitMQ server logs:

[warning] <x.xxx.x> Mnesia->Khepri fallback handling: Mnesia function failed <number> times.
Possibly an infinite retry loop; trying one last time.

The upgrade process then halts and does not complete automatically.

Environment

RabbitMQ Tile 2.3.5 → 2.4.2

Cause

The issue is triggered by the presence of the exchange logger backend configuration, which conflicts with internal Khepri DB initialization logic during the upgrade.

Problematic configuration:

{exchange, [ {enabled, true} ]}

This configuration, when present, leads to repeated Mnesia-to-Khepri fallback failures and stalls the upgrade.

Resolution

Before upgrading, remove or disable the exchange logger backend configuration from the RabbitMQ configuration file.

Remove the following block:

{exchange, [
    {enabled, true}
]}

Once removed, reattempt the upgrade from 2.3.5 to 2.4.2. The upgrade should complete without any further issues.

 

Alternative Approach

• Monit stop/start: If the upgrade is already stuck, perform the following manual steps to complete it successfully:

• • Stop the RabbitMQ server process on all nodes:
    

    monit stop rabbitmq-server

  • Start the RabbitMQ server process one node at a time: 

    monit start rabbitmq-server

• Two step Upgrade: If configuration modification is not feasible, perform a two-step upgrade instead:

• • Upgrade from 2.3.5 → 2.4.0
  • Then upgrade from 2.4.0 → 2.4.2

This approach has been observed to avoid the issue entirely.

Additional Information

Known Issue — Manual workaround available.
Future RabbitMQ Tile release will have the fix.