Support Home | Dell US
Support Home | Dell USSupport for Dell OpenManage Enterprise | Dell US
User Guide for DELL Technologies models including: OpenManage Enterprise Power Manager, Enterprise Power Manager, Power Manager, Manager
This document is only available in English.
File Info : application/pdf, 26 Pages, 1.81MB
DocumentDocumentTechnical Whitepaper Mitigate Thermal Issues in a Data Center Using OpenManage Enterprise Power Manager Abstract This technical whitepaper provides an overview about the Alert Policy feature of OpenManage Enterprise Power Manager. November 2024 Revisions Revisions Date November 2024 Description Initial release Acknowledgements Author: · Anil Kumar Maharana--Software Principal Engineer, OpenMange Enterprise Power Manager · Sreelakshmi Prakash--Software Principal Engineer, OpenMange Enterprise Power Manager Support: · Pavithra Sundar--Technical Content Developer 3, Content Engineering & Translation The information in this publication is provided "as is." Dell Inc. makes no representations or warranties of any kind with respect to the information in this publication, and specifically disclaims implied warranties of merchantability or fitness for a particular purpose. Use, copying, and distribution of any software described in this publication requires an applicable software license. Copyright © 2024 Dell Inc. or its subsidiaries. All Rights Reserved. Dell Technologies, Dell, EMC, Dell EMC and other trademarks are trademarks of Dell Inc. or its subsidiaries. Other trademarks may be trademarks of their respective owners. [012/8/2024] Technical Whitepaper 2 Mitigate Thermal Issues in a Data Center Using OpenManage Enterprise Power Manager | 635 Contents Contents 1 Introduction...................................................................................................................................................................... 5 2 Purpose of alert policies ................................................................................................................................................ 6 3 Add devices to monitoring list....................................................................................................................................... 7 4 Configure alert policy and threshold .......................................................................................................................... 10 4.1 Configure alert policy and threshold for component metrics........................................................................ 10 4.1.1 Create alert policy for component metrics .......................................................................................... 10 4.1.1.1 Configure alert policy for critical alert .............................................................................................. 10 4.1.1.2 Configure alert policy for normal alert ............................................................................................. 15 4.1.2 Configure component-level threshold.................................................................................................. 19 4.1.3 Evaluate threshold values ..................................................................................................................... 20 4.1.3.1 Threshold violation ............................................................................................................................. 20 4.1.3.2 Threshold violation recovery............................................................................................................. 22 5 Conclusion ..................................................................................................................................................................... 24 6 Appendix ........................................................................................................................................................................ 25 7 Technical support and resources............................................................................................................................... 26 7.1 Related resources .............................................................................................................................................. 26 3 Mitigate Thermal Issues in a Data Center Using OpenManage Enterprise Power Manager | 635 Executive Summary Executive Summary This technical whitepaper provides recommendations about using the Alert Policies feature of OpenManage Enterprise on the device, GPU, and CPU threshold feature of the Power Manager Plugin. Some of the key features are: · Thresholds and alerts: o Define optimal power and temperature thresholds. o Configure alert policies for metrics. · Operational actions: o Automate actions like turning on or off power based on power or temperature alerts. 4 Mitigate Thermal Issues in a Data Center Using OpenManage Enterprise Power Manager | 635 Introduction 1 Introduction Alert policy provides the flexibility to take specific actions based on alerts generated for threshold violations. These thresholds are typically set for power or temperature metrics. You can configure the various actions to be taken when an alert occurs. Use Cases: · Critical situations: For critical alerts (excessive load or overheating), immediate action (like powering off) may be necessary to prevent hardware damage. · Scheduled maintenance: During planned maintenance, customers can schedule actions like restarting servers to address noncritical alerts. Available Actions: You can configure various actions to be taken when an alert occurs: · Power off: Automatically power off the affected device or group of devices to prevent further issues. · Power on: Power on the device after addressing the alert condition. · Notify Administrators: Send notifications to relevant personnel by using an email or SMS. Also, alerts can be forwarded to other consoles. 5 Mitigate Thermal Issues in a Data Center Using OpenManage Enterprise Power Manager | 635 Purpose of Alert Policies 2 Purpose of alert policies You can perform the following actions using the Alert Policies feature when you receive alerts: · Forward alerts to other management console. · Send an SMS or email alerts. · Power off a server in response to specific error. · Power on a server in response to error recovery. 6 Mitigate Thermal Issues in a Data Center Using OpenManage Enterprise Power Manager | 635 Alert Policy Configuration 3 Add devices to monitoring list To configure threshold and apply alert policy on devices, they must be discovered in OpenManage Enterprise. After the discovery is successful, add the devices to the monitoring list in Power Manager. To discover the devices, use the OpenManage Enterprise discovery feature. Figure 1 All device page 1. Click Plugins Power Management Power Manager Devices Individual Devices Add Device. Figure 2 Add device to monitoring list 7 Mitigate Thermal Issues in a Data Center Using OpenManage Enterprise Power Manager | 635 Alert Policy Configuration 2. On the Add Devices to Power Manager page, select the device you want to add to the monitoring list and click Add Selected. Figure 3 Monitoring list page 3. To view the device metrics or to set device-level or component-level (GPU or CPU) threshold, click the name of the device on the Individual Devices page. Figure 4 Power Manager devices page 4. To view metrics, go to the Monitoring Metrics tab, and then click the Metrics tab. 5. For device metrics, click Device Metrics Trend. For component metrics--CPU and GPU--click Component Metrics Trend. 8 Mitigate Thermal Issues in a Data Center Using OpenManage Enterprise Power Manager | 635 Alert Policy Configuration Figure 5 Monitoring Metrics page 9 Mitigate Thermal Issues in a Data Center Using OpenManage Enterprise Power Manager | 635 4 Configure alert policy and threshold After a device is discovered and added to the monitoring list in Power Manager, metrics get collected for monitored devices and groups. You can configure thresholds for devices and components (CPU and GPU). This enables actions to be triggered based on the alert policies that have been configured. 4.1 Configure alert policy and threshold for component metrics 4.1.1 Create alert policy for component metrics 1. Log in to OpenManage Enterprise. 2. From the Alerts drop-down menu, select Alert Policies. 3. Click Create. Figure 6 Alert policy 4.1.1.1 Configure alert policy for critical alert On the Create Alert Policy page, do the following to create an alert policy for a critical alert on CPU or GPU temperature metrics: 1. Enter a name and description for the policy that gets triggered when critical alert is generated for GPU or CPU components. By default, the Enable Policy check box is selected to activate the policy after it is created. 10 Mitigate Thermal Issues in a Data Center Using OpenManage Enterprise Power Manager | 635 Figure 7 Alert policy name and description In the Category section, the built-in categories for OpenManage Enterprise and associated plug-ins are available. Power Manager workflows are associated with the following categories: · Application System Health Metrics · Application System Health Power Configuration · PDU Support APC, Vertiv, and Dell iPDU The Category section is generic and an optional step. Figure 8 Alert policy category 11 Mitigate Thermal Issues in a Data Center Using OpenManage Enterprise Power Manager | 635 2. To configure the alert policy for CPU or GPU temperature critical alert, enter the message ID. · For CPU, the message ID is CMET0021. · For GPU, the message ID is CMET0018. Figure 9 Alert policy message IDs 3. In the Target section, select the required target devices by clicking the Select Devices list on which this policy must be applied. Groups are not applicable for component metrics because componentlevel threshold feature is not supported for groups. Figure 10 Alert policy target 4. Select the device from the list and click OK. 12 Mitigate Thermal Issues in a Data Center Using OpenManage Enterprise Power Manager | 635 Select the target devices on the Create Alert Policy page. Figure 11 Select device for policy 5. It is recommended to use the default date and time settings for a perpetual policy. If the policy action needs to be applied only for a specific period, specify using the Date Range, Time Interval, and Days options. Figure 12 Alert policy date and time selection 6. Select the severity level as Critical for the alert. 13 Mitigate Thermal Issues in a Data Center Using OpenManage Enterprise Power Manager | 635 Figure 13 Alert policy severity selection 7. In the Actions section, select the Power Control check box, and then select Power Off from the drop-down list. This is to power off the target device when a critical alert is generated on its CPU or GPU. Figure 14 Alert policy actions The Summary section displays all the details of the alert policy to be reviewed before creation. 8. To update any values, go to the respective page and update it. 9. Click Finish to complete the Alert Policy creation for critical alert. 14 Mitigate Thermal Issues in a Data Center Using OpenManage Enterprise Power Manager | 635 Figure 15 Alert policy summary After the alert policy is created, you can edit, delete, disable, or enable the policy by selecting the respective policy. Figure 16 Alert policy page 4.1.1.2 Configure alert policy for normal alert When a threshold value for temperature metrics of CPU or GPU components returns to a normal range, the device must be powered on. To automatically power on the devices to recover, alert policy must be created for normal alert. On the Create Alert Policy page, do the following to create an alert policy for a normal alert on CPU or GPU temperature metrics: 1. Enter a name and description for the policy that gets triggered when a normal alert is generated for components. By default, the Enable Policy check box is selected to activate the policy after it is created. 15 Mitigate Thermal Issues in a Data Center Using OpenManage Enterprise Power Manager | 635 Figure 17 Alert policy name and description The Category section is generic and an optional step. 2. To configure the alert policy for normal temperature alert, enter the message ID. · For CPU, the message ID is CMET0022. · For GPU, the message ID is CMET0019. Figure 18 Alert policy message ID 3. In the Target section, select the target devices on which the critical alert policy is applied and requires a recovery. 16 Mitigate Thermal Issues in a Data Center Using OpenManage Enterprise Power Manager | 635 Figure 19 Alert policy target It is recommended to use the default date and time settings for a perpetual policy. Figure 20 Alert policy date and time selection 4. Select the severity level as Normal for the alert. 17 Mitigate Thermal Issues in a Data Center Using OpenManage Enterprise Power Manager | 635 Figure 21 Alert policy severity selection 5. In the Actions section, select the Power Control option as Power On to power on the target devices after receiving normal alert that are generated on relevant CPU or GPU components. Figure 22 Alert policy actions The Summary section displays all the details of the alert policy to be reviewed before creation. 6. To update any values, go to the respective page and update it. 7. Click Finish to complete the Alert Policy creation for normal alert. 18 Mitigate Thermal Issues in a Data Center Using OpenManage Enterprise Power Manager | 635 4.1.2 Figure 23 Alert policy summary Configure component-level threshold You can configure the temperature threshold of CPU or GPU components of a device after the policy is created. See the following example for CPU (for GPU, the steps are similar): To configure threshold for CPU, go to the Monitoring Metrics Alert Thresholds CPU Alert Thresholds tab. For GPU, go to the GPU Alert Thresholds tab. Figure 24 CPU threshold page 1. Click the Edit icon to configure the threshold value. 2. You can configure both the warning and critical values for temperature, or just one of them. You can configure the same for all the available components by selecting the Configure threshold values for all CPUs check box. 3. Click Apply. 19 Mitigate Thermal Issues in a Data Center Using OpenManage Enterprise Power Manager | 635 4.1.3 Figure 25 Configure CPU threshold Evaluate threshold values 4.1.3.1 Threshold violation 1. After configuring the value, if the CPU or GPU temperature metrics exceed the configured threshold, you receive a critical alert. Additionally, the threshold violation status can be viewed in a threshold graph in the CPU Alert Thresholds tab for CPU and in the GPU Alert Thresholds tab for GPU. Figure 26 CPU threshold graph 2. View the violation alert on the Alert Log page under the Alerts tab. 20 Mitigate Thermal Issues in a Data Center Using OpenManage Enterprise Power Manager | 635 Figure 27 Alert log page 3. After the violation becomes effective, you can log in to the iDRAC web interface for which the policy is violated and view the log which shows the shutdown message as per the policy action. Figure 28 iDRAC log data after power off View the system power state on the iDRAC page. Figure 29 iDRAC power state 21 Mitigate Thermal Issues in a Data Center Using OpenManage Enterprise Power Manager | 635 4.1.3.2 Threshold violation recovery 1. When CPU or GPU temperature metrics return to the configured threshold normal limit, you receive a normal alert. Additionally, the status gets updated in the threshold graph in the CPU Thresholds Alert tab for CPU and in the GPU Thresholds Alert tab for GPU. Figure 30 CPU threshold graph 2. View the normal alert on the Alert Log page in the Alerts tab. Figure 31 Alert log page 3. You can log in to the iDRAC web interface for which the normal alert policy is enabled and view the log which shows the power-on message as per the policy action. 22 Mitigate Thermal Issues in a Data Center Using OpenManage Enterprise Power Manager | 635 Figure 32 iDRAC log data after power on View the system power state (on) on the iDRAC page. Figure 33 iDRAC power state 23 Mitigate Thermal Issues in a Data Center Using OpenManage Enterprise Power Manager | 635 5 Conclusion Power Manager plug-in uses the Alert Monitoring feature of OpenManage Enterprise by introducing device and component-level power and temperature threshold violation alerts. It also supports reception of alerts sent by Power Distribution Unit (PDU) and Uninterruptible power supply (UPS) devices. Alert Policy feature in OpenMange Enterprise can be used to execute specific actions in response to these alerts. 24 Mitigate Thermal Issues in a Data Center Using OpenManage Enterprise Power Manager | 635 Appendix 6 Appendix Message ID CMET0008 CMET0017 CMET0004 CMET0013 CMET0015 CMET0009 CMET0018 CMET0011 CMET0020 CMET0010 CMET0019 CMET0021 CMET0022 CMET0023 Description Device power has exceeded the configured Critical threshold Device Temperature has exceeded the configured Critical threshold Device Power has exceeded the configured Warning threshold Device Temperature has exceeded the configured Warning threshold Device Power/Temperature has reached to Normal threshold GPU Power has exceeded the configured Critical threshold GPU Temperature has exceeded the configured Critical threshold GPU Power has exceeded the configured Warning threshold GPU Temperature has exceeded the configured Warning threshold GPU Power has reached the configured Normal threshold GPU Temperature has reached the configured Normal State CPU Temperature has exceeded the configured Critical threshold CPU Temperature in Normal State CPU Temperature in Warning State 25 Mitigate Thermal Issues in a Data Center Using OpenManage Enterprise Power Manager | 635 Technical support and resources 7 Technical support and resources Dell.com/support is focused on meeting customer needs with proven services and support. 7.1 Related resources · Knowledge Base for Dell OpenManage Enterprise HTML · Dell OpenManage Enterprise Power Manager Version 3.3 User's Guide PDF HTML · Dell OpenManage Enterprise Power Manager RESTful API Guide version 3.3 HTML · Dell OpenManage Enterprise Power Manager 3.3 Release Notes PDF 26 Mitigate Thermal Issues in a Data Center Using OpenManage Enterprise Power Manager | 635