As part of my home lab project, I built my servers using Supermicro X8DTH-6F motherboards for many reasons however one of these reasons was that the motherboard hosts an IPMI base management controller (BMC) interface for managing the power state of the host out-of-band along with the ability to access the IP KVM and virtual media support to allow me to access the server console session remotely and even attach .iso media over the network.
In my quest to make my lab more accessible I wanted to be able to script the procedure for starting it and shutting it down and whilst I haven’t given a whole heap of thought to the whole thing just yet, I have made an interesting discovery in the quest. Windows Server 2008 R2 introduced a PowerShell syntax PcsvDevice with various commands including Get, Start and Stop along with another Cmdlet Set-PcsvDeviceBootConfiguration.
These Cmdlets were first introduced in Windows Server 2008 R2 and the syntax for some of the parameters changed in Windows Server 2012 which is worth noting if you have used these Cmdlets previously. Using the Get-PcsvDevice Cmdlet, we can query IPMI, WS-Man and SMASH based base management controllers and get some information back from them. We need to feed this command some parameters such as a credential and an IP address for the BMC.
$Cred = Get-Credential ADMIN $Ip = "172.16.150.21" Get-PcsvDevice -TargetAddress $Ip -ManagementProtocol IPMI -Credential $Cred
As you can see, I set variables for $Cred and $Ip to set the username and the IP address to connect with and the ManagementProtocol parameter allows us to specify whether the controller is IPMI or WS-Man based. The result from this command in my environment is as follows.
So from the screenshot, we can see that the Get-PcsvDevice Cmdlet can detect my Supermicro IPMI interface and the EnabledState column in the output shows that my host is currently Disabled or in other words, powered off. So now that I know PowerShell can find my BMC, I want to be able to power on the host. By piping the Get-PcsvDevice Cmdlet into the Start-PcsvDevice Cmdlet, we can do just that.
$Cred = Get-Credential ADMIN $Ip = "172.16.150.21" Get-PcsvDevice -TargetAddress $Ip -ManagementProtocol IPMI -Credential $Cred | Start-PcsvDevice
After entering the Cmdlet, I heard my server spin up in the rack downstairs. I opened up a command prompt in a separate window and started pinging the management IP address of the host which would start pinging once Windows Server 2012 R2 started after the boot POST.
As you can see from this second screenshot, the host started to ping on its management address and I assure you, there is no magic here like me running downstairs and hitting the power button, this all happened via PowerShell remotely. Running the Get-PcsvDevice Cmdlet again now would return a similar output except the EnabledState column now reports Enabled as the host is powered on.
The opposing Cmdlet is Stop-PcsvDevice however I won’t be demonstrating this one and I’ll provide a warning to go with it. This PowerShell Cmdlet does not instruct the IPMI or WS-Man interface to perform a graceful soft shutdown using the operating system commands. It performs a hard off as it someone had just held down the power button. If your server is running and accessible via normal in-band management means, power it down that way such as the PowerShell Cmdlet Stop-Computer.
The other command which is available to us is Set-PcsvDeviceBootConfiguration. I won’t demonstrate this one either as I like my server how it’s configured right now but this Cmdlet can be used to change the boot order of a server. If for example, you wanted to set a server to Boot from LAN over PXE so that you could apply some out-of-band update or image to it then you could do that. You can get the syntax for Set-PcsvDeviceBootConfiguration or any of the other Cmdlets from TechNet at https://technet.microsoft.com/en-us/library/dn283384.aspx.