Index settings (shards, replicas) in Elasticsearch - Step-by-Step Execution

Concept Flow - Index settings (shards, replicas)

Create Index Request

↓

Set number_of_shards

↓

Set number_of_replicas

↓

Elasticsearch Distributes Shards

↓

Index Ready for Use

When creating an index, you set how many shards and replicas it has. Elasticsearch then distributes these shards across nodes.

Execution Sample

Elasticsearch

{
  "settings": {
    "number_of_shards": 3,
    "number_of_replicas": 2
  }
}

This JSON sets an index with 3 primary shards and 2 replicas for each shard.

Execution Table

Step	Action	Setting Changed	Value	Effect
1	Create index request received	none	none	Start index creation
2	Set number_of_shards	number_of_shards	3	Index will have 3 primary shards
3	Set number_of_replicas	number_of_replicas	2	Each primary shard will have 2 replicas
4	Elasticsearch distributes shards	shard allocation	3 primaries + 6 replicas	Shards spread across nodes
5	Index ready	none	none	Index can accept data and queries

💡 Index creation completes after shards and replicas are allocated

Variable Tracker

Variable	Start	After Step 2	After Step 3	Final
number_of_shards	undefined	3	3	3
number_of_replicas	undefined	undefined	2	2
total_shards	undefined	undefined	undefined	9 (3 primaries + 6 replicas)

Key Moments - 2 Insights

Why do we set number_of_shards before number_of_replicas?

Because replicas are copies of primary shards, you must first decide how many primary shards exist (see execution_table steps 2 and 3).

What happens if number_of_replicas is set to 0?

No replicas are created, so data is only on primary shards. This reduces fault tolerance (not shown in this trace but implied after step 3).

Visual Quiz - 3 Questions

Look at the execution_table, what is the value of number_of_shards after step 2?

A 3

B 2

C undefined

D 9

💡 Hint

Check the 'Setting Changed' and 'Value' columns at step 2 in execution_table

At which step does Elasticsearch distribute shards across nodes?

A Step 2

B Step 3

C Step 4

D Step 5

💡 Hint

Look for 'Elasticsearch distributes shards' in the Action column of execution_table

If number_of_replicas was set to 1 instead of 2, what would be the total shards allocated?

A 3

B 6 (3 primaries + 3 replicas)

C 6

D 9

💡 Hint

Refer to variable_tracker's total_shards calculation and adjust replicas accordingly

Concept Snapshot

Index settings control data distribution.
number_of_shards sets primary pieces.
number_of_replicas sets copies for fault tolerance.
Total shards = primaries + replicas.
Set shards first, then replicas.
Elasticsearch spreads shards across nodes.

Full Transcript

When creating an Elasticsearch index, you specify settings for number_of_shards and number_of_replicas. First, the number_of_shards is set, which determines how many primary pieces the data is split into. Then, number_of_replicas is set, which defines how many copies of each primary shard exist for safety. Elasticsearch then distributes these shards and replicas across the cluster nodes. The total shards equal the sum of primary shards and their replicas. This process ensures data is balanced and fault tolerant.