While restoring a PostgreSQL HA instance, an administrator notices that the restore of a standby node has failed. The primary node is ready to seive application traffic.
When restoring a PostgreSQL HA instance in NDB, if the standby node fails to restore but the primary node is successfully restored and ready to serve application traffic, NDB considers the database operational from the perspective of the primary node. In this scenario, NDB leaves the database in an "ONLINE" state, reflecting that the primary node is functional and capable of handling requests, even though the HA configuration is not fully intact due to the standby failure. The administrator would need to address the standby node separately (e.g., by re-provisioning or troubleshooting), but the database’s state is determined by the primary node’s readiness.
Option A (PAKFIAL.READY) appears to be a typo and is not a valid NDB state.
Option B (OFFLINE) is incorrect because the primary node is operational, so the database is not offline.
Option C (ONLINE) is correct as it reflects the primary node’s ability to serve traffic.
Option D (RECOVERY PENDING) is incorrect because the primary node has completed recovery and is active, not pending.
This behavior aligns with NDB’s focus on maintaining service availability where possible.
References
Nutanix Database Service (NDB) User Guide, Chapter 5: Configuring Time Machines, Section: Restoring HA Instances
Nutanix Support & Insights, Knowledge Base Article: "Handling Partial Restore Failures in PostgreSQL HA"
Nutanix Certified Professional - Database Automation (NCP-DB) v6.5 Blueprint, Section 5: Protect Databases Using Time Machine
Contribute your Thoughts:
Chosen Answer:
This is a voting comment (?). You can switch to a simple comment. It is better to Upvote an existing comment if you don't have anything to add.
Submit