Under normal operating conditions, the HCP nodes manage themselves. However, because hardware failures are possible, the System Management Console provides controls for shutting down and restarting individual nodes.
- Shutting down a node means shutting down the HCP software on it and powering it off. After shutting down a node, you need to manually power it back on to start HCP on it again.
- Restarting a node means shutting down and restarting the HCP software on it without powering it off.
Shutting down or restarting a node has no effect on the data in the primary storage volumes and the NFS volumes that are associated with the node. However, while a node is shut down or in the process of restarting, clients have no access to that node.
Shutting down or restarting a node is rarely necessary and can cause extra service activity, which increases the load on the system. However, situations in which shutting down a node may be appropriate include:
- A hardware failure has occurred.
- The node requires maintenance.
- You plan to remove the node from the system.
The most frequent reason for restarting a node is that your authorized HCP service provider has requested it during the assessment or resolution of a problem.
When you shut down or restart a node, you are required to specify a reason for the action.
To shut down or restart a node, you need the administrator or service role.
Shutting down an individual node
When you shut down a node, the hardware configuration determines whether it powers off automatically. To complete the action, you may need to power off the node manually after stopping it.
To ensure that HCP has enough time to shut down cleanly, wait a few minutes before powering the nodes off manually. If the node is connected to a console, the console displays System halted when you can safely power off the node.
Before you begin
Take one of the following actions:
- On the page, click the shutdown icon for the node.
- On the Storage Node page for the node, click Shut Down Node.
The Shut Down Node window appears.
In the Reason field, type the reason why you’re shutting down the node. This text can be up to 1,024 characters long and can contain any valid UTF-8 characters, including white space.
Click Shut Down Node.
The nodes shut down and, if possible, power off.
Restarting an individual node
Restarting an HCP node causes it to reboot. While the node is in the process of restarting, clients have no access to it.
When you restart a node in a system with zero-copy failover enabled, the storage managed by the node fails over to the paired node during the shutdown part of the restart. The failover process always finishes before the restarted node completes its shutdown process. Therefore, the data outage caused by the failover lasts less than three minutes.
When the node comes back up, the failback process extends the reboot by 15 to 30 seconds. The data outage caused by the failback lasts less than five minutes.
Before you begin
Take one of these actions:
- On the page, click the restart icon for the node.
- On the Storage Node page for the node, click Restart Node.
The Restart Node window appears.
In the Reason field, type the reason why you’re restarting the node. This text can be up to 1,024 characters long and can contain any valid UTF-8 characters, including white space.
Click Restart Node.
The node reboots. When the node has finished rebooting, it is again available for client access.