Droplet lifecycle
Start, stop, restart, resize, move, and delete droplets.
Droplet lifecycle actions change guest execution or resource configuration. Each action returns an operation whose outcome must be checked before starting a dependent change.
Start and stop
Open Compute → Droplets → Droplet → Actions for lifecycle controls. Hardware changes are under Hardware; protection is under Configuration. The action review links to its operation after submission.
Start schedules and boots a stopped droplet. Placement can change if its previous node is unavailable and its volumes support another eligible node. Required node or zone constraints still apply.
Shut down asks the guest to stop through the guest agent or hypervisor power interface. The operation waits for the configured grace period. If the guest does not stop, the operation reports that condition; it does not silently turn a graceful request into a forced power-off.
Force stop powers off guest execution without waiting for the operating system. Buffered writes can be lost and filesystems can require recovery. Use it only when the application impact is acceptable.
Start requires a stopped or eligible preempted guest and available placement. Shutdown and restart require running execution. Graceful shutdown waits for guest acknowledgement; force stop does not.
Resize requires a stopped droplet. It changes configured CPU and memory, not volume contents. Deletion requires no conflicting operation and disabled deletion protection. Each action preserves or removes dependencies according to its own workflow, so a stopped guest and a deleted guest have different retention effects.
Resize
Create a recovery point, stop the droplet, and select Resize. Choose a size available to its region and pool. Infra checks quota, image requirements, CPU profile, and node capacity before applying the configuration.
The action does not resize filesystems or data volumes. A failed admission check leaves the previous size unchanged. If the existing node cannot hold the larger droplet, select a compatible destination through the resize preview. Start the droplet after the operation succeeds and verify application health.
Cold moves
A cold move transfers a stopped droplet when live migration is unsupported. The preview lists local disks that must be copied and devices that must be reassigned. The source remains unavailable during the copy and destination boot.
The destination must be in the same region and satisfy CPU, device, storage, and placement requirements. A failed copy retains the source data until the operation determines which artifacts are safe to remove. Do not delete the source manually while the operation is reconciling.
Deletion protection
Enable Deletion protection on production droplets to reject delete requests from the console and API. An operator with configuration permission must first disable protection. The change is recorded in the audit log.
Protection does not prevent shutdown, guest disk corruption, or loss of the underlying hardware. It is not a substitute for backups.
Delete a droplet
Open Delete and review each dependency. The root volume is selected for deletion by default when it was created with the droplet. Separately managed data volumes and reserved public addresses are retained unless explicitly selected and eligible for removal. Backup retention is governed by its policy.
Deletion stops execution before removing interfaces and selected disks. Failed
cleanup leaves the operation with retained resource IDs. Repeating a delete
against an already removed ID returns 404; use operation history to determine
whether the original deletion succeeded.
There is no droplet recycle bin. Restore retained disks or backups into a new droplet through Disaster recovery.
Scratch and local data
A normal stop retains persistent volume attachments but discards ephemeral scratch contents. Local persistent disks retain their contents on their source node and require a supported copy before another placement can use them.
Review local and ephemeral disks before stopping a worker for maintenance. A cold move can copy a persistent local disk while creating empty scratch storage at the target.