This page covers shutting down and powering on the system, and updating the SC//HyperCore™ HyperVisor software.
Shut Down the System
The system should always be shut down from the SC//HyperCore web interface for any planned network or infrastructure management that may impact power or network availability to the nodes. Always contact Scale Computing Support to review your individual scenario if you would like to shut down individual nodes.
- Open the Control tab in the Control Center.
- Under System, click Shutdown. A confirmation dialog will appear.
- Select Yes or No from the Force dropdown menu. Selecting Yes will forcibly power off any powered-on VMs during the shutdown process.
- Click Shutdown Cluster to confirm and shut down the system.
Power On the System
All system nodes should be powered on as closely together as possible. Any significant delay in powering on one or more nodes will trigger a recovery scan and prolong normal startup checks. The system will go through a 20–30 minute startup check process. Only conditions that remain after the initial wait period should be addressed.
Update the System
* Updates will always be disruptive for Single Node Systems, as there is no VM failover destination.
SC//HyperCore software updates are an important part of regular system maintenance. Almost all SC//HyperCore updates are rolling, meaning they are non-disruptive to running VMs as long as there is sufficient RAM to live migrate VMs between nodes while a node is updated.
When an update is available, it will appear in the Update tab in the Control Center and as a flag above the current SC//HyperCore version in the HUD.
- Open the Update tab in the Control Center.
- If an update is available, the update version and release date will be displayed.
- Click Apply Update.
- Click Update to confirm and begin the update.
The update process will work through each node sequentially. VMs running on the node being updated will be live migrated to nodes with free resources. If there is insufficient RAM to live migrate the VMs, the system will display an alert and abort the update. Sufficient RAM must be made available before starting the update again.
Once the update on a node is complete, VMs will be moved back to that node and the update process will proceed to the next node until all nodes have been updated.