Should You disable keepReplicationJournalImage on Read-Only replicating Clusters?

Yes, you can choose not to set keepReplicationJournalImage = true on your consumption (read-only) clusters.

Since these clusters are not intended to function as replication journal (RJ) nodes and are never expected to take over RJ responsibilities, there is no advantage in enabling this setting. Doing so would introduce unnecessary overhead by making the cluster store a local copy of the central journal data, which it will never use for failover or replication purposes.

You are following the correct approach by enabling this setting only on your active or standby RJ clusters, where it serves a clear and necessary purpose:
- Creating a redundant journal image
- Supporting manual failover in disaster recovery scenarios

Therefore, it is completely safe and appropriate to leave this setting disabled on consumption clusters.

Behavior of keepReplicationJournalImage = true in a Multi-Cluster Setup

Assume you have three clusters — Cluster 1, Cluster 2, and Cluster 3 — with the replication journal hosted on Cluster 1. Clusters 2 and 3 are configured to use the replication journal on Cluster 1 via the cn=replicationjournal data source.

If you enable keepReplicationJournalImage = true on Cluster 2:

Write Behavior:
Cluster 2 continues to write all its changes to the central journal on Cluster 1. In addition, it stores a backup of all journal entries it receives (including those from Cluster 3) into its own local cn=replicationjournal, creating a replica of the central journal.

Read Behavior:
This setting does not alter how Cluster 2 reads updates. It still fetches journal entries only from Cluster 1 and does not use its local backup for reading.

Disaster Recovery Use Case:
The local copy on Cluster 2 can be used as a manual failover option if Cluster 1 experiences a prolonged outage. In such a scenario, support must manually update the system to point to the backup journal. When Cluster 1 is restored, another manual switch is needed to revert operations back if desired.

Why This Matters:
This setup ensures that critical RJ clusters maintain a recent backup of the central journal. However, enabling this setting on additional clusters may lead to increased resource usage, as each cluster must write all changes locally.

Best Practice:
Enable keepReplicationJournalImage only on clusters that may act as failover candidates for the central journal — typically your active or standby RJ clusters. There is no need to enable it on consumption or read-only clusters.