Intel,Altiris Group

Integrating Intel AMT Power Control Into Everyday Operations 

May 12, 2010 01:12 PM

Much has been posted about the goodness that Intel vPro Technology provides, specifically integrated management engine within Intel AMT (Active Management Technology). This enables out-of-band manageability providing capabilities like reliable power control, boot redirection, KVM remote control, and more.

With all the features and demonstrations that have been shown, a basic question keeps coming up - "How do I use Intel AMT to power-on my clients at a designated time?"

The nature of the question might vary - yet the core of the question is similar. In various customer and training experiences, some may have misused the remote power control capabilities of the technology. When an Intel AMT power-off command is sent - don't expect a graceful action. It's a literal change from an "on" power-state to an "off" power-state. "Why?" you might ask - well, keep in mind - this is out-of-band management focus. The intent was for situation when a client has stopped responding, where an agent based or operating system command is not affecting the client yet you simply need it powered off or rebooted.

Similar inquiries have focused on using Intel AMT in combination with power policy management solutions, combining the technology with rigorous Green IT solutions which can provide electric utility quality reports for energy rebates, and so forth. Keep in mind - Intel AMT is an enabler to the end solution... not the other way around.

Let me share an example how to better manage the power within your own environment using just Symantec Management Platform and Intel vPro Technology.

Get Power Scheme Task Agent - Last I checked, this is a free download\install. It helps to enforce the Windows Power Scheme on the clients. This does require the Power Scheme Task sub-agent to be installed on the target clients. Using TaskServer, a defined power scheme policy can be scheduled and delivered on a recurring basis thus helping to ensure a policy is enforced within the environment.

Take a look at the two screenshots below

Next - if a graceful power-off task or routine is needed, focus on an agent or operating system based approach. Have you ever tried running "shutdown -f -s" on your client? (in case you do and want to reverse it, quickly type "shutdown -a"). What about a WMI call to initiate a power down? Or - what about a Power Control Task delivered via TaskServer policy as a schedule local client job?

Take a look at the example below

Now you might be wondering - where does Intel AMT play into all of this?

Intel AMT can be used to power-on the clients when needed - instead of using a Wake-on-LAN broadcast message. As with any tool or capability - do some testing and validation on what works within your environment. If curious on what might happen when hundreds or thousands of clients power-on within a few minutes of each other - take a look at a previous posting - What Happens When a Few Hundred Systems Power-on via vPro?

The golden rules testing, "your mileage may vary", and so forth still apply. The technology will respond as directed - how that response affects the environment (including the Notification Server's ability to send commands to all target clients) should be consideration.

To schedule a 1-to-many power-on event, take a look at the following:

First - define a power-on task using the Real-Time Console Infrastructure. These are the Intel AMT tasks sent from the server to the client

The single task can be used as needed - targeting collections or individual systems.

The single task can be combined with other tasks. For example - power on via Intel AMT, deliver a software package, and then power-off.

Side comment here - if you've enabled the Power Scheme Task with setting for clients to go to sleep after 10 minutes of inactivity, followed by hibernate\off after 1hour - will that suffice? If not, include the client task for graceful shutdown.

The example TaskServer job shows a three step sequence as previously mentioned.

For some customer environments - they simply need a power-on event at designated times. One example I've come across is an environment where updates\maintenance from various applications occur at defined intervals during off-hours. For that customer, a power-on event via Intel AMT is scheduled to occur at 11pm and 1am every weekday. The reason is that the maintenance job is expected to occur - from different systems than the Symantec Management Platform - at 11:10pm and 1:10am. The Windows Power Policies on the clients will put the client's to sleep after 20 minutes of inactivity (refer to the first item - Power Scheme Task Management to enforce power policies). In this scenario - all the TaskServer job needs to do is power-on. (Okay - for those Workflow orchestration experts out there - you might have some commentary about how this could all be combined into one process with dependencies\conditionals\etc. Let's face it - not every customer needs a full blown workflow solution for these types of tasks. They are willing to piecemeal the solution for now... BUT - if you have a really cool example, I'm very interested to hear it)

Once the tasks have been defined - whether to enforce a Power Scheme Policy or sequence a power\patch job - the next step is to define targets and schedules.

The screens below provide examples of scheduling a job to run every weekday at 1am local time, along with the target clients of the job.

Within your own environment - do some testing to ensure the target clients are indeed powering on and completing the other tasks.

One method to check the success\failure status is by selecting the target job. In the example below - if you double click, a status window will appear and provide the ability to drill down on the overall and system specific status in completing each of the tasks in he overall job.

I'm interested to hear about your own experiences on integrating Intel AMT power control within your everyday operations.

The opinions expressed on this site are mine alone and do not necessarily reflect the opinions or strategies of Intel Corporation or its worldwide subsidiaries

Statistics
0 Favorited
0 Views
0 Files
0 Shares
0 Downloads

Tags and Keywords

Comments

Sep 14, 2010 05:06 PM

There's a follow-on article with additional information.   See http://www.symantec.com/connect/articles/optimizing-rtci-intel-vpro-operations

Information shared comes from both lab and real-world customer experiences.   Expect more insights on optimization, tuning, key learnings, and so forth.  And as always - I welcome hearing from the community on your own experiences.... good and bad.

May 24, 2010 05:49 PM

Hopefully a successful outcome on the revisit

May 24, 2010 05:47 PM

Curious to know if the "TLS issues" reference is to remote configuration certificates or TLS certificates in the configuration profile.

In general - if you do not currently encrypted traffic on your internal network, it may be better to forego use of TLS in the configuration profile.    It may ease the setup\configuration and ongoing maintenance of the platform.

If TLS is required - curious to know what types of issues you are facing

May 20, 2010 12:21 AM

Thanks for sharing!  I too need to revisit OOB as well and look forward to trying this out immediately for the power savings.

May 19, 2010 12:56 AM

Great article. I need to re-visit our OOB setup, I had TLS issues in the past...

Related Entries and Links

No Related Resource entered.