Resource Actions

VM Actions Explained¶

Once you have a VM workload running, you'll find several actions available in the Resources table under the VirtualMachine resource. Here's what each one does:

Action	What It Does	When to Use It
Stop	Immediately terminates the VM and releases compute resources back to the cluster. This is equivalent to yanking the power cord.	When you're done with the VM and want to free up resources. Unsaved work will be lost.
Restart	Fully deletes the existing VM instance and recreates it from scratch. This is a hard reset—yank the plug and plug it back in.	When the VM is unresponsive or you need a clean reboot from the hypervisor level.
Soft Reboot	Sends a reboot signal to the guest OS (Windows). This is the same as clicking "Restart" from within Windows.	For routine reboots — after installing updates, applications, or when Windows asks for a restart.
Pause	Freezes the VM in its current state. The VM remains in memory but stops executing.	Special use cases only — such as temporarily freeing CPU cycles or troubleshooting. Resume when ready.
Migrate	An offline migration moves a virtual machine from one node to another by shutting it down on the source node and restarting it on the target node without maintaining uptime	when you need to move a virtual machine between nodes for maintenance, resource optimization, or troubleshooting and can tolerate downtime.
Evict	Sets a custom node affinity that evicts the VM off its current node, so it is rescheduled onto a different node.	When the current node needs to be cleared, for example ahead of node maintenance, and you don't need to pick the destination node.
Clear Affinity	Clears the custom node affinities set by the Evict and Migrate actions. Only available once one of those affinities has been set. See Custom Affinities.	When the VM no longer needs to be kept off, or pinned to, a particular node.

Tip

For day-to-day reboots, use Soft Reboot — it's graceful and lets Windows shut down properly. Use Restart only when the VM is stuck or unresponsive. Stop is for when you're truly done and want resources back.

Deployment/StatefulSet Actions Explained¶

Deployment and StatefulSet resources support hibernating, migrating, and evicting an individual resource without deleting the workload.

Action	What It Does	When to Use It
Hibernate	Scales the resource's replicas to 0, deleting its running pods and freeing up the compute those pods were using.	When you want to free up compute for a resource you aren't actively using, without deleting the workload entirely.
Activate	Scales a hibernated resource's replicas back up, recreating its pods.	When you're ready to resume using a hibernated resource.
Migrate	Sets a custom node affinity that moves the resource from its current node to another node, rescheduling its pods there.	When you need to move a resource between nodes for maintenance, resource optimization, or troubleshooting.
Evict	Sets a custom node affinity that evicts the resource off its current node, so its pods are rescheduled onto a different node.	When the current node needs to be cleared, for example ahead of node maintenance.
Clear Affinity	Clears the custom node affinities set by the Evict and Migrate actions. Only available once one of those affinities has been set. See Custom Affinities.	When the resource no longer needs to be kept off, or pinned to, a particular node.

Migrating and evicting restart the resource's pods

Moving a resource to another node recreates its pods, so anything living only in a pod's filesystem or memory is lost. To persist data across a move, mount a Persistent Volume Claim (PVC) to the workload.

Hibernating does not save session state

Scaling to 0 replicas deletes the pod, so anything living only in the pod's filesystem or memory is lost. To persist data across a hibernate/activate cycle, mount a Persistent Volume Claim (PVC) to the workload.

Note

Only Deployment and StatefulSet resources support hibernation. Workloads made up of other resource kinds will not show a hibernate action.

Custom Affinities¶

The Migrate, Evict, and Clear Affinity actions are available on VirtualMachine, Deployment, and StatefulSet resources.

Admin only

Migrate, Evict, and Clear Affinity are restricted to admin users. Other users will not see these actions in the resources table.

The Evict and Migrate actions work by setting a custom node affinity on the resource. Orion tracks these custom affinities, so it knows which placement rules were added by an action rather than defined by the workload's template.

While a resource has a custom affinity set, the Clear Affinity action becomes available. Clearing removes only the custom affinities set by Evict and Migrate, leaving any affinities from the workload's template in place, so the resource is free to be scheduled on any eligible node again.

Service Actions Explained¶

Service resources have a connection info action available.

Action	What It Does	When to Use It
Connection Info	Returns the service's fully qualified domain name (FQDN) and port.	When you need to reach the service directly, for example from another workload in the cluster or when configuring a client application to point at it.

Tip

Workloads exposed through a LoadBalancer ingress also surface their ingress endpoints in the workload's CONNECT dropdown, see LoadBalancer Ingress Endpoints.

DataVolume Actions Explained¶

DataVolume resources have a clone action available.

Action	What It Does	When to Use It
Clone	Creates an exact copy of the volume, including its data and storage configuration. This will then be tracked by Orion and shown in the Genesi Data Volumes table. It will also be available to add to a workload template from within Genesis.	Use this if you want to utilize this Data Volume for another workload. Example, when creating a golden image for a VM workload