Frequently Asked Questions on Migration

Q: How do I revert from RTFS back to Legacy Federation?

A: If the Artifactory Federation Service (RTFS) is enabled (rtfs.enabled = true) and you have migrated, either manually or through auto-migration, you must use the RTFS Migration Tool to revert to Legacy Federation. Manually setting rtfs.enabled = false without running the tool first results in an inconsistent state. For the full rollback procedure, see Migration Rollback.

Q: What happens if I disable rtfs.enabled without using the migration tool?

A: Manually changing the rtfs.enabled configuration flag without using the migration tool can result in:

• Inconsistent Federation state between your configuration and actual queue status
• System alerts in the Artifactory logs indicating a mismatch between configuration and queue status
• Synchronization failures between Federation members
• Potential data inconsistency across Federated repositories

Always use the migration tool for any RTFS state changes.

Q: What is auto-migration and when does it occur?

A: Auto-migration is an automated process that completes the RTFS migration status. It doesn't automatically migrate existing Legacy Federation deployments to RTFS. Auto-migration only occurs when both conditions are met:

1. RTFS is already enabled (rtfs.enabled = true)
2. You have no existing Federated repositories

In this scenario, the system automatically sets the migration status to MIGRATION_COMPLETED since there are no repositories requiring migration.

Important: If you are currently using Legacy Federation with existing Federated repositories, auto-migration won't migrate you to RTFS automatically. You must use the RTFS Migration Tool to perform the migration.

By default, the auto-migration job is enabled and runs every 4 hours.

To disable auto-migration, add the following to your system.yaml:

artifactory:
  rtfs:
    auto:
      migration:
        enabled: false

Q: What are the migration status states?

A: The migration process goes through the following states.

To check your current migration status:

curl -u <USERNAME>:<PASSWORD> -X GET \
  "https://<JFROG_URL>/artifactory/api/federation/migration/status"

Where:

• <USERNAME> and <PASSWORD>: Your Artifactory login credentials
• <JFROG_URL>: The base URL of your JFrog Platform instance

Q: Do I need to run the migration tool on every HA node?

A: No. You only need to run the migration tool against one node per JFrog Platform Deployment (JPD). The Artifactory node that handles the request applies the RTFS change and then attempts to notify the other HA nodes. However, this notification can fail, for example during a rolling upgrade. When it does, the health-check job doesn't automatically re-enable or align RTFS on the lagging nodes, so you may need to manually remediate them by retrying the migration or restarting the lagging nodes. See Enable or Disable RTFS in an HA Cluster for the full behavior and the log messages to look for. You must still run the migration tool on each JPD in your Federation.

Q: Can I run RTFS and Legacy Federation simultaneously during migration?

A: Yes. During the transition period, Hybrid Mode allows RTFS and Legacy Federation to coexist across different sites. Hybrid Mode operates at the site level, not the repository level, which means:

• One JPD can run on RTFS
• Another JPD in the same Federation can continue running on Legacy Federation
• You can migrate your sites incrementally, one JPD at a time

Hybrid Mode doesn't allow mixing RTFS and Legacy Federation within the same site. Once a JPD is migrated to RTFS, all Federated repositories on that JPD use RTFS.

This approach allows you to:

• Test RTFS on a non-critical site first
• Validate functionality before migrating production sites
• Maintain Federation connectivity between migrated and non-migrated sites during transition

Hybrid Mode is intended as a transition state, not a permanent configuration.

Q: Does reverting to Legacy Federation require downtime?

A: No. The migration tool is designed to perform the rollback without requiring downtime. However, during the rollback process:

• Federation synchronization may be temporarily paused
• Some events may be queued until the rollback completes
• All JPDs in the Federation should be accessible