Failover Clustering and Network Load Balancing Team Blog
This blog will discuss the proper process for shutting down an individual node in a Windows Server 2012 R2 Failover Cluster or the entire cluster with all the nodes. Note: While the steps outlined will be specific to Windows Server 2012 R2, the process applies to a cluster of any OS version.
When shutting down or rebooting a node in a Failover Cluster, you first want to drain (move off) any roles running on that server (such as a virtual machine). This ensures that the shutting down of a node is graceful to any applications running on that node.
When the node is once again powered on and ready to be put back into service, use the Resume action to re-enable the node to host roles.
PS C:\> Resume-ClusterNode –Failback Immediate
Shutting down the entire cluster involves stopping all roles and then stopping the Cluster Service on all nodes. While you can shut down each node in the cluster individually, using the cluster UI will ensure the shutdown is done gracefully.
When the cluster is shut down, the VMs will be placed in a Saved state. This can be controlled using the OfflineAction property.
To configure the shut down action for an individual VM (where "Virtual Machine" is the name of the VM):
PS C:\> Get-ClusterResource "Virtual Machine" | Set-ClusterParameter OfflineAction 1
The VM is turned off
The VM is saved
The guest OS is shut down
The guest OS is shut down forcibly
Thanks! Elden Christensen Principal Program Manager Lead Clustering & High-Availability Microsoft
I've a quick Question...Can you please explain technical differences between drain (move off) and manual Failover of all the individual services? For example, Let's say..I've a Clustered SQL Server Instance, what's difference does it make from SQL Server perspective if I failover SQL services to the otehr node and reboot the node vs Drain Roles and reboot the node?
Putting a node into maintenance mode by invoking the node drain is a management paradigm. Technically under the covers it exactly the same thing as going and manually discovering all the groups on a node and then manually issuing a group move. Drain just makes it easier... it's a single click action. But either way works, so do what you prefer.
Thanks Elden for the clarification...
Thanks Elden for sharing the same ...
Does shutting an hyper-v cluster stop all running VMs?
Hi Elden Christensen,
Thanks for good Artical about How to Properly Shutdown a Failover Cluster or a Node.
I am planning to learn from scrarch Could you please provide the PDF file on Clustering on Windows 2008 server.
Please find my personal Mail ID: email@example.com.
Thanks & Reagrds,
Just wanted to echo Antonio's question
If I shutdown the whole cluster (mine is full of Hyper-V VMs) what happens to the wirtual machines?
Do they save their state or us SCVMM or native Hyper-V rules?
Is there a GUI process in failing over a Windows 2008 cluster to other node?Just like how it was handled in Windows 2003 clusteramdin?
Is there a best practice for bringing a cluster back online?
I have an Windows 2012R2 failover cluster and the Active Directory was made high available due to the fact that from 2012 the cluster can boot-up without the AD. I tested this in case a hard-reset the physical machines and it works.
But if I manually shutdown the cluster, how can I start-up a cluster back again? Because I can start-up the cluster if I'm not authenticated on a Domain.
I only see the start command via powershell. Is it possible to start via failover cluster manager?