Create a Couchbase dSource using XDCR

Use XDCR when you want continuous replication of Couchbase data to a staging cluster. XDCR keeps the staging database synchronized with the source.

Before you begin

• Source cluster is accessible from staging host over network

• Source cluster admin credentials with required privileges:

  • XDCR_ADMIN – Required to set up XDCR replication

  • DATA_MONITOR – Required to monitor cluster and replication status

• Staging cluster admin credentials (for creating new staging cluster)

• Network connectivity:

  • Source to staging: SSH (22), Couchbase management (8091), XDCR replication (11210)

Procedure

1. Navigate to Data > Data Management > dSources.

2. Select + dSource

3. Select:

   1. Source-based dSource

   2. Couchbase source config

4. Configure ingestion settings

   1. dSource type: XDCR

   2. Mount path – Base path for NFS mounts (e.g., /mnt/provision). Must be empty and writable by environment OS user.

   3. Staging host – FQDN recommended for cluster communication

   4. Staging port – Default 8091

5. Define resource allocation for staging cluster creation:

   1. Cluster RAM size – Total cluster memory

   2. Index RAM size – Memory for indexing service

   3. FTS RAM size – Memory for full-text search (if needed)

   4. Eventing RAM = 0 – Disable on staging (not required for replication)

   5. Analytics RAM = 0 – Disable on staging (optional; enable on target VDBs if needed)

6. Configure authentication

   1. Staging cluster admin username/password.

   2. Source host – FQDN recommended for cluster communication.

   3. Source port – Default 8091.

   4. Source cluster admin username/password (must have XDCR_ADMIN and DATA_MONITOR privileges).

7. Configure bucket settings

   1. Bucket size:

      • 0 – Inherit bucket size from source cluster (recommended)

      • Any positive integer (MB) – Override all buckets to use this size

   2. Optional: Specify bucket list for selective replication (replicate only specific buckets)

8. Complete wizard configuration

   1. dSource name

   2. Target group

   3. Environment user

9. Configure optional settings:

   1. Snapshot policy

   2. Hooks (optional)

10. Select Submit

During XDCR linking

The Continuous Data Engine performs the following operations:

• Validates mount path accessibility

• Verifies credentials and privileges on both clusters

• Creates staging cluster with specified resource allocation

• Configures XDCR replication from source to staging

• Starts data synchronization

Result

A dSource is created from the source cluster and is ready for snapshots and VDB provisioning.

If linking fails

If dSource linking encounters errors, check the following:

• XDCR specific:

  • Verify source cluster credentials have XDCR_ADMIN and DATA_MONITOR privileges

  • Confirm network connectivity on port 11210 (XDCR binary protocol)

• Common issues:

  • Mount path inaccessible or not empty

  • SSH access issues to staging host

  • Couchbase binaries missing or wrong version on staging

  • Port conflicts or firewall blocking connections

Review Continuous Data Engine logs and Couchbase instance logs at /opt/couchbase/var/lib/couchbase/logs on staging host for detailed error messages.