Mirrored queues for redundancy in RabbitMQ - Step-by-Step Execution

Process Flow - Mirrored queues for redundancy

Client sends message

↓

Message arrives at Master Queue

↓

Master replicates message

↓

Messages copied to Mirror Queues

↓

Consumers read from Master or Mirror

↓

If Master fails -> Mirror promoted to Master

↓

Message flow continues without loss

Messages sent to the master queue are copied to mirror queues on other nodes. If the master fails, a mirror takes over to keep messages safe.

Execution Sample

RabbitMQ

rabbitmqctl set_policy ha-all "" '{"ha-mode":"all"}'
rabbitmqctl list_queues name policy

# Publish message to queue 'task_queue'
# Consumer reads from queue
# Master failure triggers mirror takeover

This setup command creates mirrored queues across all nodes for redundancy.

Process Table

Step	Action	Master Queue State	Mirror Queues State	Result
1	Create queue 'task_queue' with ha-mode=all	Empty	Empty on all mirrors	Queue created and mirrored
2	Publish message 'msg1' to master	Contains 'msg1'	Contains 'msg1' replicated	Message replicated to mirrors
3	Consumer reads 'msg1' from master	Empty	Contains 'msg1'	Message acknowledged on master
4	Master node fails	Unavailable	Contains 'msg1' on mirrors	Mirror promoted to master
5	Consumer reads 'msg1' from new master	Empty	Empty after ack	Message consumed without loss
6	Publish message 'msg2' to new master	Contains 'msg2'	Contains 'msg2' on mirrors	New messages replicated
7	Consumer reads 'msg2' from new master	Empty	Empty after ack	Message consumed
8	Master restored and rejoins cluster	Empty	Empty	Cluster stable with mirrored queues

💡 Execution stops after cluster stabilizes with mirrored queues ensuring no message loss despite master failure.

Status Tracker

Variable	Start	After Step 2	After Step 3	After Step 4	After Step 5	After Step 6	After Step 7	Final
Master Queue Messages	Empty	['msg1']	Empty	Unavailable	Empty	['msg2']	Empty	Empty
Mirror Queues Messages	Empty	['msg1']	['msg1']	['msg1']	Empty	['msg2']	Empty	Empty
Master Node Status	Active	Active	Active	Failed	Failed	Active	Active	Restored
Mirror Node Status	Active	Active	Active	Active (promoted)	Active (promoted)	Active (promoted)	Active (promoted)	Active

Key Moments - 3 Insights

Why does the message still exist on mirror queues after the master queue message is acknowledged?

Because the master acknowledges the message first, but mirrors keep a copy until they also confirm consumption, ensuring no message loss if master fails (see step 3 and 5).

What happens to message flow when the master node fails?

The mirror queue is promoted to master and continues message delivery without interruption, as shown in step 4 and 5.

Why do we need to set the policy with ha-mode=all?

This policy tells RabbitMQ to mirror the queue on all nodes, ensuring redundancy and failover capability (step 1).

Visual Quiz - 3 Questions

Look at the execution table, what is the state of the master queue after step 3?

A Empty

B Contains 'msg1'

C Unavailable

D Contains 'msg2'

💡 Hint

Check the 'Master Queue State' column at step 3 in the execution table.

At which step does the mirror queue get promoted to master?

A Step 2

B Step 4

C Step 6

D Step 8

💡 Hint

Look for 'Mirror promoted to master' in the 'Result' column.

If the ha-mode policy was set to 'nodes' instead of 'all', how would the message replication change?

A Messages replicate to all nodes

B No replication happens

C Messages replicate only to specified nodes

D Messages replicate only to master

💡 Hint

Refer to the explanation in key moments about ha-mode policy effects.

Concept Snapshot

Mirrored queues replicate messages across nodes for redundancy.
Set policy with ha-mode=all to mirror on all nodes.
Master queue handles messages; mirrors keep copies.
If master fails, a mirror is promoted to master.
This ensures no message loss and high availability.

Full Transcript

Mirrored queues in RabbitMQ work by replicating messages from a master queue to mirror queues on other nodes. When a client sends a message, it arrives at the master queue and is copied to mirrors. Consumers read from the master queue normally. If the master node fails, one of the mirrors is promoted to master, allowing message consumption to continue without loss. This setup requires setting a policy with ha-mode=all to enable mirroring on all nodes. The execution trace shows queue states and message flow through publishing, consumption, failure, and recovery steps, illustrating how redundancy is maintained.