Node Drain Management
Node draining allows providers and admins to gracefully take nodes offline for maintenance. When a node is drained, it is hidden from the marketplace and active users are notified, ensuring a smooth transition before the node goes offline.
Drain Workflow
1. Start Draining
Begin the drain process on a node. This hides the node from the marketplace so no new instances are scheduled on it, and notifies current users.
POST /api/v1/nodes/drain
2. Wait for Instances to Terminate
Allow existing instances to finish their work or be migrated. Users are notified and can take action to move their workloads.
3. Force Drain (Optional)
If instances do not terminate within an acceptable timeframe, force-drain the node to immediately terminate all remaining instances.
POST /api/v1/nodes/force_drain
4. Perform Maintenance
With all instances cleared, perform your hardware or software maintenance.
5. Resume the Node
Bring the node back online and make it available on the marketplace again.
POST /api/v1/nodes/resume
Cancelling a Drain
If maintenance is no longer needed, cancel an active drain to restore the node to normal operation without completing the full drain cycle.
POST /api/v1/nodes/cancel_drain
Only providers and admins can perform drain operations on nodes in their datacenters.
Node drain management is available through the CloudRift Console and REST API. See the API documentation for full details.