PowerEdge: Power Settings

What do the different power supply settings within the iDRAC mean?

Table of Content

1. Part 1 - Where to Find Power Supply Unit (PSU) related Information?
2. Part 2 - Changing Power Settings
3. Part 3 - Common Troubleshooting Scenarios

Part 1 - Where to Find Power Supply Unit (PSU) related Information?
There are two places to look within the iDRAC web Graphical User Interface (UI).

1. System > Overview > Power 
   This section gives a lot of historical data about the server's power consumption and PSU status, output wattage, and input line type. 

In particular, know that the "Input Line Type" and "Output Wattage" can be important to troubleshooting. Several 14G and later power supplies automatically detect if they are plugged into high line power (~220 V) or low line power (~110 V) outlets and swap to the appropriate output wattage. Sometimes, being on low line power limits how much wattage a PSU can output. In this example, the PSUs are rated at 1100 W, but are only capable of outputting 1050 W when connected to low line power. Other PSU examples can be more extreme, like a 2000W PSU only being able to output 1000W of power in a low line power environment. 

Occasionally, customers move their servers to a new office location without realizing that they are moving from a high line power environment to one with low line power. This results in the server no longer being able to power on. The iDRAC would still be reachable in this scenario using Direct-Connect and support can come to this page and use the information in the above graph as well as the other information about this page in the web UI to help diagnose the customer's problem. 

2. Configuration > Power Management 

This page provides us with the power settings that we can switch, including powering the server on or off, PSU redundancy, and more.  
 

Part 2 - Changing Power Settings 

Let us go through each setting: 

1. Power Control

This setting is exactly as it sounds - you can power the server on and off with this setting. You can perform a graceful shutdown, a warm or cold reboot, or power the server off. You can also send a Non-Maskable Interrupt (NMI) to an installed, non-Linux OS to force it to stop. This can be used for diagnostics and should only be used for specific troubleshooting or debugging scenarios. 

2. Power Cap Policy 

This is an iDRAC Enterprise or Datacenter licensed feature, so it does not appear on all iDRACs. It allows server administrators to set the maximum AC power load that the server is allowed to request from the data center. The server automatically adjusts its performance to remain below the power cap, but it can temporarily exceed this cap under moments of high demand. 

3. Power Configuration

Probably the most commonly used settings, the Power Configuration section allows admins to set up PSU redundancy, hot spares, and power factor correction (PFC). 

PSU Redundancy has three possible settings. 

• Not Redundant - Every PSU contributes to the server's power. This can allow the server to reach its maximum possible power draw, but the lack of redundancy also means that if one PSU fails, the server can experience an outage. 
• A/B Grid Redundant - Power supplies are divided into Grid A and Grid B. If a grid or a PSU on one grid fails, but the PSUs on the second grid are functional, the system does not shut down.
• PSU Redundant - Similar to A/B Grid Redundant, except that the PSUs are not grouped into grids. Instead, each PSU is an individual contributor to the server's power, allowing for more redundancy. This option is not available in every server. 

Hot Spare 

• Enabling a hot spare allows admins to set one or more PSUs as 'standby' allowing them to enter a low-power state until they are needed as part of a failover scenario. This puts almost all the load onto the 'active' PSUs, which may not be desirable for some clients. If a customer needs the load split evenly among all their PSUs, then it is best to disable the hot spare feature. 

Primary PSU 

• This setting is only applicable if the "Hot Spare" setting is enabled. The primary PSU allows admins to determine which PSU is the 'active' PSU upon which most of the power demand is placed. 

Power Factor Correction 

• Power Factor Correction (PFC) allows the PSUs in the server to operate more efficiently and reduce the amount of 'wasted' power. In large data centers, large amounts of wasted power can lead to fines and penalties by service providers. Alternatively, disabling PFC allows the PSUs to drain when the server is in the "S5" power state (AKA: Powered off, but still plugged in). A deep understanding of the customer's power needs is necessary to determine if this setting should be disabled, and so support generally does not advise on whether enabling or disabling this setting is appropriate. 

Part 3 - Common Troubleshooting Scenarios 

There are two common power-related troubleshooting scenarios that engineers are faced with: 

1. The server no longer powers on after moving to a new location
2. The server no longer powers on after installing new hardware

In both scenarios mentioned support may be able to provide temporary relief by disabling PSU redundancy. 

For the first scenario, if the customer moved from a high line power environment to a low line power environment, then the maximum output wattage of a single PSU may no longer be enough to meet the server's power demands. Disabling PSU redundancy allows the server to draw from additional power supplies, thus increasing the maximum amount of power that can be supplied to the server. 

The second scenario is similar to the first, except the maximum output of a single PSU is unchanged. Instead, the customer may have installed hardware that pushes the server's power needs beyond that of a single PSU. Again, disabling PSU redundancy can allow the server to come online. If so, the customer may have to speak to their sales representative about purchasing PSUs that can support their new power needs while maintaining redundancy. 

YouTube version of the video: