Skip to main content

Set Up Replication

Replications copy Iceberg tables — metadata and data files — from a source catalog to a target catalog.


Create a Replication

  1. Navigate to Iceberg > Replications
  2. Click Create New Replication
  3. Fill in the form:
FieldDescription
NameA descriptive label for this replication
Source catalogThe catalog to replicate from
Target catalogThe catalog to replicate to (cannot be the same as source)
ModeContinuous or One-time (see below)
Copy typeLatest snapshot or All snapshots (see below)
Inclusion scopesOne or more scopes selecting which tables to replicate
Exclusion scopesOptional scopes for tables to skip
  1. Click Create

Choose a Mode

ModeWhen to use
ContinuousKeep the target in sync as changes happen. Requires a Hive source catalog (uses the Metastore event stream). Runs until stopped.
One-timePerform a single sync pass. Works with all catalog types. Use for initial loads, dev snapshots, or ad-hoc copies.

If the source catalog is not Hive, continuous mode is not available and the mode is automatically set to one-time.


Choose a Copy Strategy

Copy typeBehaviour
Latest snapshotCopies only the most recent table state. Faster and uses less storage, but the target will not have snapshot history.
All snapshotsCopies the complete snapshot history. Preserves time-travel and rollback capabilities on the target. Use for disaster recovery.

Add Location Mappings

If the source and target catalogs use different storage locations, you must create a location mapping so Ice Flow knows where to write data files.

  1. Go to the Replications page and click the Location Mappings tab
  2. Click Create New Location Mapping
  3. Select the Source warehouse and enter the Source path
  4. Select the Target warehouse and enter the Target path
  5. Click Create

The mapping is now available to associate with any replication via the replication's Mappings tab.

See Manage Warehouses for more on warehouse and mapping management.


Start and Stop a Replication

  • A newly created replication starts in the Inactive state
  • Use the Start action to begin replication
  • Use Stop to pause a continuous replication
  • One-time replications transition to Complete automatically

Replication States

StateMeaning
InactiveCreated but not yet started
ReplicatingActively copying data (continuous mode: watching for changes)
StoppedManually stopped or paused
CompleteOne-time replication finished all work
EnforcedPaused by license enforcement

Monitor Replication Progress

Operations Tab

The Operations tab shows table-level sync events. Each row represents one table synchronised during one replication cycle:

ColumnDescription
IdentifierUnique operation ID
NamespaceNamespace containing the table
TableTable name
TypeOperation type (COMMIT, CREATE, etc.)
Start TimeWhen the operation started
End TimeWhen the operation completed
Data FilesNumber of data files committed to the target
Total DataBytes transferred
DurationTime taken

Operations are paginated. Use the controls to navigate large result sets.

File Transfers

Click an operation row to reveal the File Transfers panel. This shows each file copied during the operation, with its size and transfer time. A chart visualises file sizes and durations. File transfers are also paginated.

Scopes Tab

Read-only view of the inclusion and exclusion scopes configured for this replication.

Mappings Tab

Shows which location mappings are associated with this replication. Click a mapping to view or edit it.

Settings Tab

Edit the replication name, notes, or delete the replication.


Edit a Replication

  1. Open the replication from Iceberg > Replications
  2. Click the Settings tab
  3. Modify the name or notes
  4. Click Save

To change scopes, mode, or catalogs, delete the replication and create a new one.


Delete a Replication

  1. Stop the replication if it is running
  2. Open the Settings tab
  3. Click Delete Replication