Replica set architecture mental model in MongoDB - Step-by-Step Execution

Concept Flow - Replica set architecture mental model

Client sends request

↓

Primary node receives

↓

Primary applies write

↓

Primary replicates to secondaries

↓

Secondaries apply write

↓

Secondaries acknowledge

↓

Primary confirms write success

↓

Client receives confirmation

↓

If primary fails

↓

Election starts

↓

New primary elected

↓

Clients redirect to new primary

The client sends requests to the primary node, which applies writes and replicates them to secondary nodes. If the primary fails, an election chooses a new primary to keep the system running.

Execution Sample

MongoDB

Client -> Primary: Write data
Primary -> Secondaries: Replicate data
Secondaries -> Primary: Acknowledge
Primary -> Client: Confirm write
If Primary fails -> Election -> New Primary

Shows the flow of a write operation and failover in a MongoDB replica set.

Execution Table

Step	Action	Node Involved	State Change	Result
1	Client sends write request	Client	Request sent to Primary	Primary receives write
2	Primary applies write	Primary	Data written to Primary	Write successful on Primary
3	Primary replicates write	Primary -> Secondaries	Write sent to Secondaries	Secondaries receive write
4	Secondaries apply write	Secondaries	Data written to Secondaries	Secondaries updated
5	Secondaries acknowledge	Secondaries -> Primary	Acknowledgement sent	Primary receives acknowledgements
6	Primary confirms write	Primary -> Client	Confirmation sent	Client receives success
7	Primary fails	Primary	Primary goes down	No response from Primary
8	Election starts	Secondaries	Vote for new Primary	One Secondary elected Primary
9	New Primary elected	New Primary	Becomes Primary	Clients redirect writes here

💡 Execution stops after new primary is elected and clients redirect their requests.

Variable Tracker

Variable	Start	After Step 2	After Step 4	After Step 6	After Step 7	After Step 9
Primary status	Up	Up	Up	Up	Down	Up
Secondaries status	Up	Up	Up	Up	Up	Up
Write data state	Not written	Written on Primary	Written on Secondaries	Confirmed to Client	Confirmed	Confirmed
Client write status	Not sent	Sent	Waiting	Success	Waiting for new Primary	Redirected

Key Moments - 3 Insights

Why does the client send writes only to the primary node?

Because the primary node is the only one that accepts writes and replicates them to secondaries, as shown in execution_table rows 1 and 2.

What happens when the primary node fails during operation?

An election starts among secondaries to choose a new primary, as shown in execution_table rows 7 to 9, ensuring continuous availability.

How do secondaries confirm they have applied the write?

They send acknowledgements back to the primary, which then confirms success to the client, as shown in execution_table rows 4 and 5.

Visual Quiz - 3 Questions

Look at the execution_table, at which step does the client receive confirmation that the write succeeded?

A Step 5

B Step 3

C Step 6

D Step 9

💡 Hint

Check the 'Result' column for when the client receives success confirmation.

According to variable_tracker, what is the status of the primary after step 7?

A Down

B Up

C Unknown

D Recovering

💡 Hint

Look at the 'Primary status' row after step 7.

If the primary fails, what is the next action in the replica set flow?

A Client retries write to primary

B Secondaries start election

C Secondaries shut down

D Client disconnects

💡 Hint

Refer to execution_table steps 7 and 8 for failover process.

Concept Snapshot

Replica set has one primary and multiple secondaries.
Client sends writes only to primary.
Primary applies writes and replicates to secondaries.
Secondaries acknowledge writes.
If primary fails, secondaries hold election.
New primary elected to maintain availability.

Full Transcript

In a MongoDB replica set, the client sends write requests to the primary node. The primary applies the write and then replicates it to the secondary nodes. The secondaries apply the write and send acknowledgements back to the primary. Once the primary receives acknowledgements, it confirms success to the client. If the primary node fails, the secondaries start an election to choose a new primary. After a new primary is elected, clients redirect their writes to the new primary, ensuring continuous operation.