Type
  • Best Practice
Description
  • This article explains the procedure to follow to monitor Dell PowerEdge servers with Hardware Sentry.
Additional Keywords
  • Dell, PowerEdge

Related Topics

Knowledge Base

Monitoring Dell PowerEdge Servers with Hardware Sentry

KB1039 - Sep 16, 2010 - Last reviewed on May 16, 2019

Contents

This document covers the Dell PowerEdge server series running either Microsoft Windows or any supported distribution of Linux.

About Dell PowerEdge Servers

There is a single main server series in Dell’s portfolio: PowerEdge. Dell PowerEdge servers are standard PC servers based on X86 or X64 processors (from Intel or AMD). These servers can run either Microsoft Windows (any server version supported by Microsoft) or Linux (RedHat, SuSe and many other distributions). The operating system type has no impact on the hardware monitoring solution.

Most Dell PowerEdge servers feature an internal RAID disk controller: a Dell PERC controller.

Note on Dell pre-installed with Sun Solaris

Dell allows its customers to purchase PowerEdge servers that are shipped with Sun Solaris pre-installed. The impact on the hardware instrumentation layer is yet unknown.

Hardware Instrumentation

SNMP

Hardware Sentry uses SNMP to communicate with Dell OpenManage Server Administrator. In most cases, SNMP is installed as part of the operating system installation. If not, the SNMP service/snmpd daemon must be installed, configured, and started on your Dell PowerEdge server before installing Dell OpenManage Server Administrator.

In-band: Dell OpenManage Server Administrator

To monitor the hardware of Dell PowerEdge servers, you need to install Dell OpenManage Server Administrator (OMSA) on your Dell PowerEdge server.
This hardware agent covers all hardware aspects of Dell PowerEdge servers: temperatures, fans, power supplies, controllers, physical disks, processors, memory modules, etc.
A good way to check if Dell OMSA is properly installed and configured is to connect to its Web interface. Typically, the URL to connect to is https://‹hostname›:1311/

Dell OpenManage Server Administrator can be downloaded from Dell’s Website.
Dell OpenManage Server Administrator - Web Interface

Out-of-band: Dell Remote Access Card (DRAC or iDRAC)

Some Dell PowerEdge servers are equipped with an out-of-band management card (DRAC or iDRAC) that allows administrators to manage and monitor their servers remotely even when no operating system is running. On some old systems, the SNMP connectivity of the DRAC is weak and does not provide much hardware information. Therefore, Hardware Sentry cannot rely on DRAC to discover and monitor Dell PowerEdge servers. However, any Dell server with a model number with a middle number of 2 or 3 (e.g.: PowerEdge R720xd or R730) has an iDRAC7 or later (G12 / G13) and should be monitored via the iDRAC, especially if it is running ESX.

IPMI

Most recent Dell PowerEdge servers use the IPMI standard for their internal instrumentation. The motherboard and/or the DRAC card features a BMC chip (Baseboard Management Controller) which can be accessed either internally or over the network through the IPMI protocol (if the BMC chip has been configured so). IPMI only provides environment monitoring (temperatures, voltages, fans, etc.) and does not expose any information regarding the disk controllers, physical and logical disks.

Hardware Sentry does not use IPMI directly to monitor Dell PowerEdge servers. It relies on the Dell OpenManage Server Administrator agent, which in turn relies on the IPMI technology for some of its operations (like sensing temperatures etc.)

Setting up Hardware Sentry on Dell PowerEdge Servers

Pre-requisites

  • The Dell PowerEdge server runs Microsoft Windows or a supported distribution of Linux
  • SNMP is running on the server (Windows SNMP service, or snmpd daemon on Linux). SNMP versions 1, 2c, and 3 (which needs a valid username/password) must be supported at least in read-only mode.
  • SNMP is to be configured to allow requests from local host with at least a community string enabled in read mode.

Installation procedure

  • Install on the server the version of the Dell OpenManage Server Administrator agent corresponding to the Dell PowerEdge server (it depends on the operating system, its version and the processor architecture).
  • Install, on the same server, the PATROL Agent (v.9.x.xx minimum is recommended)
  • and the Hardware Sentry KM (v.1.9.50 or higher).
  • Install Hardware Sentry on the server (this can be done at the same time as the PATROL Agent). Please follow the installation instructions of the Hardware Sentry User Guide.

Configuration on Microsoft Windows systems

On Microsoft Windows systems, there is no specific configuration required for Hardware Sentry to discover and monitor the hardware of Dell PowerEdge servers.

Configuration on Linux systems

On Linux systems, the SNMP community configured in the snmpd daemon may not be “public”. In such a case, it is essential to verify that the SNMP community version selected to communicate with the SNMP agent installed on the monitored host is correct. (see Hardware Sentry User Guide).

The monitoring of the network cards does not rely on Dell’s agent but on certain Linux command line utilities:

  • ifconfig
  • ethtool
  • mii-tool

Root privileges required

Both ethtool and mii-tool require root privileges. This means that Hardware Sentry needs to be configured to either execute commands as root or to use sudo for the ethtool and mii-tool programs. Sudo avoids the need to provide the KM with the password or the root account, but it requires the proper configuration of the /etc/sudoers file with your root account. The monitoring of the network cards will not work properly until the root account credentials are provided or the KM has been configured to use sudo.

Configuring Hardware Sentry KM for PATROL

The below instructions explain how to configure Hardware Sentry KM for PATROL to remotely monitor a Dell PowerEdge server:

In the PATROL Console

To configure the remote monitoring of a Dell PowerEdge server:

  1. Right-click the Hardware icon and select KM Commands > Add a Remote System or an External Device...

    Adding a Remote Host — System Identification**

  2. Identify the Dell PowerEdge server you want to monitor remotely:

    • System or Device name: Enter the name of the Dell PowerEdge server you wish to monitor. In our example: “madrid” (a PowerEdge 1850 server).
    • IP address or fully qualified name: Enter the IP address or the fully qualified name of the Dell PowerEdge server you wish to monitor. You can leave this field blank if the IP address can be resolved by your DNS/hosts file.
    • From the System or Device type drop-down list, select the system type used by your Dell PowerEdge server. In our example: “Microsoft Windows”.
  3. Click Next.

    Adding a Remote Host — Connector Selection Mode**

  4. Specify the connector selection mode:

    • (Default and Recommended) Automatically detect the suitable connectors based on: In this mode, Hardware Sentry will run tests at each discovery and select the connectors that best match your Dell PowerEdge server.
    • (Advanced) Manually choose which connectors to use: With this mode, you will have to manually select the connectors Hardware Sentry will use to monitor your Dell PowerEdge server.
    • If you have previously selected the “automatic” option:
      1. Select the protocol(s) to be tested by Hardware Sentry. In our example: “SNMP”.
      2. Click Next.
    • If you have previously selected the “manual” option:
      1. Click Next.
      2. Select the connectors Hardware Sentry will use to monitor the Dell PowerEdge server. You can refer to the Hardware Connector Library documentation to find out which connectors are suitable for your server.
      3. Click Next.
  5. You now need to configure the Connection Settings for the SNMP protocol that will be used by Hardware Sentry to collect information.

    Adding a Remote Host — SNMP Connection Settings

  6. Select the SNMP version to be used and click Next.

  7. If you have previously selected SNMP version 1, indicate:
    • the Community to be used.
    • the Port number (default: 161).
  8. If you have previously selected SNMP version 2c, indicate:
    • the Community to be used.
    • the Port number (default: 161).
    • the number of seconds Hardware Sentry will wait for an SNMP response (Default timeout: 120 seconds). This timeout covers all the sub-queries that compose the parent query.
  9. If you have previously selected SNMP version 3, indicate:
    • The Username to be used to perform the SNMP query.
    • The Authentication protocol to be used to authenticate the SNMP v3 messages. Possible values are: None, MD5, SHA.
    • The Authentication password to be used to authenticate the SNMP v3 messages.
    • The Privacy protocol to be used to authenticate SNMP v3 messages. Possible values are: None, AES, DES.
    • The Privacy password associated with the privacy protocol.
    • The Context name accessible to the SNMP entity.
    • The Port number (default: 161).
    • The number of seconds Hardware Sentry will wait for an SNMP response (Default timeout: 120 seconds). This timeout covers all the sub-queries that compose the parent query.
  10. Click Next.
  11. Nothing else needs to be configured. Click Finish.

Hardware Sentry will perform a new full discovery to detect your Dell PowerEdge server. An icon labeled Hardware on < remote hostname > will appear. It represents the server you are now monitoring remotely, along with icons for all of the detected components of this remote server.

inline

In the TrueSight Operations Management Console

To configure the remote monitoring of a Dell PowerEdge server in CMA:

  1. Create a new policy or edit an existing one.

    Adding a Remote Host — Device Information

  2. In the Device Information section, identify the Dell PowerEdge server you want to monitor remotely:

    • Device Name: Enter the name of the Dell PowerEdge server you wish to monitor.
    • IP Address: Enter theIP address of your Dell PowerEdge server. You can leave this field blank if you specified the host name in the Device Name field and if the IP address can be resolved by your DNS/hosts file.
  3. From the Device type drop-down list, select the system type used by your Dell PowerEdge server.

    Configuring the Protocol/Connection  Information

  4. In the Protocol/Connection Information section, click SNMP.

  5. Define the SNMP Information:
    • Select the version of the SNMP protocol used by the device to be monitored. Possible values are 1, 2c, or 3.
    • Use the spin button to indicate the SNMP port number you wish to use to perform SNMP queries. By default the SNMP queries are performed through port 161.
  6. If you selected SNMP Version 1, indicate the SNMP Community string to use to perform SNMP v1 queries.
  7. If you selected SNMP Version 2c, indicate:
    • the SNMP Community string to use to perform SNMP v2c queries.
    • the number of seconds Hardware Sentry will wait for an SNMP response. By default, the timeout is set to 120 seconds to cover all the sub-queries that compose the “parent"queries.
  8. If you selected SNMP Version 3, indicate:
    • The Username to be used to perform the SNMP v3 queries.
    • The Authentication Protocol to be used to authenticate the SNMP v3 messages. Possible values are: None, MD5, SHA.
    • The Authentication password to be used to authenticate the SNMP v3 messages.
    • The Privacy protocol to be used to authenticate SNMP v3 messages. Possible values are: None, AES, DES.
    • The Privacy password associated with the privacy protocol.
    • The Context name accessible to the SNMP entity.
    • The Port number (default: 161).
    • The number of seconds Hardware Sentrywill wait for an SNMP response. By default, the timeout is set to 120 seconds to cover all the sub-queries that compose the 'parent' queries.
  9. Click Close to validate.
  10. Nothing else needs to be configured. Click the Add to List buttonto validate and add the defined device to the list below.
  11. In the Add Monitor Types dialog box, click Addor Update.

Discovered Components and Monitored Parameters

Connectors

When configured properly, the following connectors should be automatically selected by Hardware Sentry to monitor a Dell PowerEdge server:

On Windows:

  • Dell OpenManage Server Administrator
  • Dell OpenManage Storage Manager
  • WMI – Network or MIB-2 Standard SNMP Agent - Network Interfaces - Windows

On Linux servers:

  • Dell OpenManage Server Administrator
  • Dell OpenManage Storage Manager
  • Linux – Network or MIB-2 Standard SNMP Agent - Network Interfaces - Linux

Components and monitored parameters:

The following components and parameters are discovered and monitored:

  • Server model
  • Per-chassis overall status
  • Intrusion in a chassis (if opened when powered on)
  • Temperature sensors: actual temperature and status
  • Fans: speed and status of each fan
  • Voltage sensors: actual voltage and status
  • Power supplies: status
  • Memory modules: size and type, status
  • Processors: type and frequency, status
  • DRAC card status (if any)
  • Status of the Embedded System Management Log (ESM Log)
  • RAID disk controllers, firmware version, battery status (if any), controller status
  • Physical disks: vendor, size, serial number, and status
  • Logical disks (RAID volumes): size, RAID type, and status
  • Network cards: vendor, model, connection speed, status, link status and error percentage

Troubleshooting

Failure to communicate through SNMP

On Windows:

If Hardware Sentry fails to automatically detects the SNMP community that is configured and allowed in the Windows SNMP Service, you can configure Hardware Sentry to use a specific SNMP community to make its requests (see Hardware Sentry’s User Guide).

If the PATROL Agent’s default account does not have administrative privileges, it is likely that it will not be able to perform any WMI requests. Hardware Sentry uses WMI to monitor the network cards of Dell PowerEdge servers. In such case, you would either need to give administrative rights to the PATROL Agent’s default account, or configure the WMI service to accept connections and requests from the PATROL account. This can be done through the DCOMCNFG utility.

If Hardware Sentry fails to automatically detect the Dell OpenManage Server Administrator SNMP agents (Server Administrator and Storage Manager), it might be because:

  • The PATROL Agent is configured to disable SNMP communications (SNMP as a client, or a “manager”, but not as an agent). Please check the /snmp/support configuration variable in the PATROL Agent’s configuration.
  • The Windows SNMP service is neither installed, nor configured, or nor started. It is not installed by default in Windows 2003 and Windows 2008. When installed, it is configured by default to reject any SNMP request, from any host, from any community.
  • The Dell OpenManager Server Administrator agent failed to install correctly. Please try to uninstall and re-install it.

On Linux:

If Hardware Sentry fails to automatically detect the Dell OpenManage Server Administrator SNMP agents (Server Administrator and Storage Manager), it might be because:

  • The PATROL Agent has been configured to disable SNMP communications (SNMP as a client, or a “manager”, but not as an agent). Please check the /snmp/support configuration variable in the PATROL Agent’s configuration.
  • The snmpd daemon is neither installed, nor configured, or nor started.
  • Hardware Sentry has not been properly configured to use the correct SNMP community string. Check the snmpd configuration files and configure Hardware Sentry to use the proper SNMP community string. Refer to the Configuring the SNMP Connection Settings section of the Hardware Sentry User Guide for further details.
  • The Dell OpenManage Server Administrator agent failed to install correctly. Please try to uninstall and re-install it.
  • The Dell OpenManage Server Administrator SNMP sub-agent failed to register with the snmpd SNMP master agent because of a conflict between snmpd and the PATROL Agent’s own SNMP master agent: snmpmagt. In this case, it is strongly recommended to disable the SNMP master agent of the PATROL Agent. Set the /snmp/agent_auto_start configuration variable to “no” in the PATROL Agent’s configuration. Disable “SNMPStart” parameter in Linux.km.

Duplicate instances

Duplicate instances can be observed for each environment sensor (temperature, fan, voltage, and power supply) when Hardware Sentry is configured to execute external commands with the root account while the system is instrumented with an IPMI BMC chip. This problem results from the detection of both the Dell OpenManager Server Administrator SNMP agent and the ipmitool utility as valid hardware information sources. When this situation occurs, remove the ipmitool connector. Using the Hardware Sentry interface:

In the PATROL Console

  1. Under Detected connectors,right-click the ipmitool connector › KM Commands › Remove.

In the TrueSight Operations Management Console

  1. Edit your existing policy.
  2. In the Device Advanced Settings section, click Connectors.
  3. Select the Exclude Specified Connectors option from the Connector Selection list.
  4. Enter MS_HW_IpmiTool.hdf in the Connector File Names field.
  5. Click Close to validate.