Indexing and replication separation in indexer clusters

Indexing and Replication Separation decouples indexing from bucket replication in Splunk Enterprise indexer clusters by streaming hot bucket slices through SmartStore-backed remote storage.

Field Value
Release Splunk Enterprise 10.4 and higher
Applies to Classic indexer clusters
Requires Ingest-Tier Scaling
Storage SmartStore-enabled indexes

How Indexing and Replication Separation works

In a traditional indexer cluster, an indexer both writes incoming data and replicates bucket data to other peers to satisfy the configured replication factor. Indexing and Replication Separation changes that behavior for configured indexes.

When you enable the feature on an index:

  • The source peer still creates and owns the hot bucket.
  • The bucket is marked as a streaming source.
  • Peer indexers are assigned as streaming target peers.
  • Hot bucket slices are streamed through SmartStore-backed remote storage instead of following only the traditional hot bucket replication path.

Replication work for configured indexes is no longer handled only through the original inline peer-to-peer hot bucket replication flow.

Benefits

  • Less indexing backpressure: Reduces replication-related backpressure on indexing.
  • Cleaner separation: Improves isolation between indexing work and replication work.
  • Better long-term architecture: Simplifies the long-term clustered indexing architecture.

Indexing and Replication Separation extends the separation-of-concerns model introduced by Ingest-Tier Scaling.

Requirements for indexing and replication separation

Review the system requirements, prerequisites, and replication factor rules before you enable Indexing and Replication Separation.

System requirements

To enable Indexing and Replication Separation, you need the following:

  • Splunk Enterprise 10.4 or higher
  • A classic Splunk indexer cluster
  • Replication already configured and functioning correctly
  • Ingest-Tier Scaling already enabled
  • SmartStore configured for every index that uses the feature

Prerequisite: Ingest-Tier Scaling

Ingest-Tier Scaling must already be enabled before you enable Indexing and Replication Separation.

Important: Do not attempt to enable Indexing and Replication Separation before you enable Ingest-Tier Scaling. This feature depends on the architecture that Ingest-Tier Scaling introduces. If Ingest-Tier Scaling is not already configured and stable, complete that deployment first.

Prerequisite: SmartStore

SmartStore must already be configured for every index where you plan to enable this feature.

If an index does not have SmartStore enabled, this feature does not work for that index. The feature is enabled per index, not globally.

Replication factor and search factor requirements

  • For single-site deployments, the replication factor (RF) must be greater than 1.
  • For single-site deployments, the search factor (SF) must equal the RF.
  • For multisite deployments, configure the RF to equal the site count, with one copy in each site.