Configure Hardware Sentry KM

Hardware Sentry monitors the hardware of almost any system locally or remotely. This section provides general information about the configuration properties of monitored systems via the Monitoring Studio Web console.

Configuring Systems Monitoring

Hardware Sentry is designed to automatically detect the components of the server it is running on along with the PATROL Agent. However, specific properties may need to be configured to ensure the product collects and reports data and performance as needed.

Step 1 - Adding a System (Host)

To add a new system (host) to your monitoring environment

  1. Click KMs > Hardware Sentry from the Web Console
  2. Click the New System… button
  3. Configure the following properties:

A. Localhost Monitoring

No specific configuration is generally required as Hardware Sentry automatically detects the components of the server it is running on along with the PATROL Agent. However, if the localhost has been removed (<Delete>) from the monitoring environment, you need to add it manually:

To add the localhost to the monitoring environment, provide the following information:

Property Description
Hostname or IP Address Enter the name of the hostname of the local system, localhost, its fully qualified domain name (FQDN), or IP address.
Associate to a Specific FQDN (Optional) Enter the FQDN to associate the system to another device in TrueSight. The FQDN uniquely distinguishes a device from any other system and ensures that a device monitored via their management cards or through controllers, for example, is properly identified.
Device Type Automatically set to match the localhost operating system (AIX, Linux, HP-UX, Windows, etc.) or type of the device (Network, Storage, etc.).
Internal ID ID to be used to store the system configuration in the PATROL Agent configuration. This Internal ID is automatically generated based on the Hostname or IP Address provided but can be edited if needed.

B. Remote System Monitoring

Remote monitoring allows you to monitor multiple hosts from a single Agent. This feature will prove to be useful if you lack resources or time to deploy a PATROL Agent and Hardware Sentry on several systems.

Hardware Sentry offers the possibility to add remote hosts to your monitoring environment individually or in a batch when the hosts share the same configuration settings (OS, protocol, credentials, etc.). When creating a group of hosts, you are proposed to define the group name (Internal ID) as well as the set of hosts included into a group being created. The name (Internal ID) makes it easier for you to identify the hosts that are part of a group in the Web console.

Additionally, you can associate your system to Fully Qualified Domain Names (FQDNs) of specific devices in TrueSight when you define the connection settings of your host(s). The FQDN designates the specific location of a device within the Domain Name System (DNS) hierarchy; it communicates the host’s position relative to the root of the DNS namespace. An FQDN enables each connected device to be uniquely identified and located within the monitored environment and therefore ensures that a device is properly identified.

Remote monitoring is required if you need to:

  • Monitor the shared chassis of a group of blade servers
  • Monitor the server itself through its out-of-band management card (which provides more information than the operating system in-band)
  • Monitor a SAN switch, disk array or tape library, which the PATROL Agent and KM cannot be installed on.

Adding a remote host or an external device to your monitored environment consists in identifying the host(s)/device(s) you wish to monitor and configure the connection settings according to the protocol you need to use. This section provides step-by-step instructions for adding a single host/device or for adding multiple hosts/devices at a time, that will be easily identified as part of a group in the Web console.

To add remote system(s) to the monitoring environment, provide the following information:

Single System

Provide the following information for each host you wish to monitor:

Property Description
Hostname or IP Address Enter the Hostname, fully qualified domain name (FQDN) or IP address of the remote system or device you wish to monitor.
Associate to a Specific FQDN (Optional) Enter the FQDN to associate the system to another device in TrueSight. The FQDN uniquely distinguishes a device from any other system and ensures that a device monitored via their management cards or through controllers, for example, is properly identified.
Device Type From the drop-down list, select the System or device to monitor.
Internal ID ID to be used to store the system configuration in the PATROL Agent configuration. This Internal ID is automatically generated based on the Hostname or IP Address provided but can be edited if needed.

Multiple Systems

It is possible to group hosts that share the same characteristics (device type, protocols, credentials, etc.) with a single system configuration. To create a host group you can use one of the methods described below:

Method 1: Adding Hosts Individually

Provide the following information for each host you wish to monitor:

Property Description
Hostname or IP Address Enter the Hostname, fully qualified domain name (FQDN) or IP address of the system or device you wish to monitor.
Associate to a Specific FQDN (Optional) Enter the FQDN to associate the system to another device in TrueSight. The FQDN uniquely distinguishes a device from any other system and ensures that a device monitored via their management cards or through controllers, for example, is properly identified.

Tip: Use the Add Host button, to keep adding hosts to an existing group.

Method 2: Adding Hosts in Bulk

Click the Bulk Add button and select a local file containing hosts details in one of the following formats (one host per line):

# Comments are ignored
# Hostname TrueSight Device FQDN
node01-clusterA clusterA.acme.com
172.16.8.200 clusterA.acme.com

# Hostnames Only
node01-clusterA
172.16.8.200

# Hostname,TrueSight Device FQDN
node01-clusterA,clusterA.acme.com
Method 3: Adding Hosts via an External File

Alternatively, you can use an external file that contains a list of the hosts you want to monitor. The file must be installed on the server hosting the PATROL Agent and its path must be entered in the Hostname or IP address field, prefixed by @ (example: @C:\myHostFile.txt). The file must follow one of these formats (one host per line), where the second column for TrueSight Device FQDN is optional:

node01-clusterA;clusterA.acme.com
node01-clusterA
172.16.8.200
Property Description
Device Type From the drop-down list, select the System or device common to all hosts specified above.
Internal ID ID to be used to store the host group configuration in the PATROL Agent configuration. This Internal ID is automatically generated based on the first Hostname or IP Address provided but can be edited if needed.

Each specified hostname will be monitored independently in the Web console and in TrueSight with a separate MS_HW_MAIN instance. They will be grouped under a MS_HW_HOSTGROUP instance. This feature is particularly useful when monitoring lots of similar hosts.

The same hostname/IP address may be used across multiple hosts. FQDNs are used to create the devices in TrueSight. FQDNs are optional.

Tip: Click the Check button next to the hostname to verify that the system resolves correctly on the DNS.

Step 2 -Selecting the Connection Protocol(s)

In order to access a system, at least one supported protocol should be configured. Hardware Sentry will choose the connectors that use this protocol to monitor the hardware components on this system.

Field Description
HTTP Enable this protocol to use HTTP/HTTPS for monitoring the hardware on a system. You will then have to specify the credentials and encryption protocol to access the system.
OS Commands Enable this protocol to use OS commands for monitoring the hardware on local systems. You will then have to specify the credentials to access the system.
SNMP Enable this protocol to use SNMP for monitoring the hardware on a system. You will then have to select the SNMP version and its corresponding settings.
WBEM Enable this protocol to use WBEM for monitoring the hardware on a system. You will then have to specify the credentials and settings to access the system.
WMI Enable this protocol to use WMI for monitoring the hardware on a Windows system. You will then have to specify the credentials to access the system.
SSH Enable this protocol to use SSH commands for monitoring the hardware on remote systems. You will then have to specify the credentials to access the system.
Cisco UCS API Enable this protocol to use Cisco UCS API for monitoring the hardware on a system. You will then have to specify the credentials and encryption protocol to access the system.
IPMI-Over-LAN Enable this protocol to use IPMI-over-LAN for monitoring the hardware on a system. You will then have to specify the credentials to access the system.

Step 3 - Specifying the Connectors to Use

Hardware Sentry comes with a library of connectors dedicated to different hardware instrumentation layers. The connectors are automatically selected based on the device type and enabled protocols.

Field Description
Connector Selection Method of selecting the suitable connectors for this system (Automatic, Use the Connectors Specified Below, or Exclude the Connectors Specified Below).
Available Connectors Selection of connectors available for use (when Use the Connectors Specified Below is selected), or exclusion (when Exclude the Connectors Specified Below is selected).

Click the SAVE button at the top of the page to save the changes to the PATROL Agent. Once saved, the corresponding MS_HW_MAIN instance will appear in the Web console. If the configured system contains several hostnames, several instances will be created under a MS_HW_HOSTGROUP instance.

Additional Settings

Monitoring network cards communication performance may be critical to ensure that they run at the expected speed in order to avoid network congestion or any other performance issue. By default, Hardware Sentry triggers alerts when a break in the network link is detected for the following parameters:

  • the Link Status parameter triggers a WARNING if a network interface previously connected to the network is detected unplugged. However, it does not trigger any alert for network interfaces that have never been connected.
  • the Duplex Mode parameter triggers a WARNING when an adapter that was running in “full-duplex” suddenly starts communicating in “half-duplex”.
  • the Link Speed parameter triggers a WARNING when the adapter “slows down”, i.e. re-negotiates the link speed with its remote counterpart, and is forced to downgrade to a slower communication rate.

To change the default settings, select the type of alert Hardware Sentry should trigger (INFO, WARN or ALARM) or simply disable the functionality (OFF).

Power Wattage

Hardware Sentry lets you set an approximation of the wattage used by a system when no power consumption data is available. Specify the electricity consumption in Watts for the system to have Hardware Sentry provide you with an estimated energy cost.

KM Settings

The KM Settings page enables you to define thresholds and severity alert levels, choose notification delivery methods and specify the troubleshooting or recovery actions you want to implement when a problem is detected on a monitored system.

Alert Settings

You can define the way Hardware Sentry notifies you when alert conditions are detected on a monitored parameter. The alert content can include macros that will be replaced at runtime.

Alert Actions Description
Annotation To display a message at the annotation point of the parameter graph.
PATROL Event To customize PATROL Event types and related event content.

  • Hardware Problems: To select the type of event to trigger upon hardware failures - Hardware Sentry Specific (MS_HW_MAIN/HardwareProblem) or Generic (STD/41) and customize the event content

  • Connector Problems: To select the type of event to trigger upon connector failures - Hardware Sentry Specific (MS_HW_MAIN/ConnectorProblem) or Generic (STD/41) and customize the event content
Command Line To execute a command line on the system where the PATROL Agent is installed.
PSL Script To specify the PSL statement to be executed locally by the Agent.
Email To send an email to one or multiple recipients.
Write to a Log File To add a user-defined entry to the Log file.
Send an SNMP Trap To send an SNMP Trap.

Customizing Alert Content

Hardware Sentry allows you to customize the content of Alert Actions with macros when a hardware failure is detected. The following macros can be used to obtain more details about the problem. They will be replaced at runtime.

Macros Description
%{/variable_name} For advanced users only. Value of the variable_name instance variable. Example: %{/worstParam} will give the name of the “worst parameter” of the instance that triggered the alert..
%{ALARM_TYPE} Type of alert triggered (ALARM, WARNING or INFORMATION).
%{ASCTIME:...} Current date and time formatted as specified in the macro. Example: %{ASCTIME:%m %d %T %Y} will produce Aug 25 11:14:53 2018. The available formats for the %{ASCTIME:…} macro are listed in the Format Symbols for %{ASCTIME:…} macros section of this documentation.
%{CONSEQUENCE} Description of the possible consequence of the detected problem. Example: The temperature of the chip, component or device that was cooled by this fan should grow quickly. This can lead to severe hardware damage and system crashes.
%{DATE} Current date in the YYYY-MM-DD format.
%{FULLREPORT} Full hardware health report about the instance that triggered the alert.
%{HOSTNAME} Name of the computer monitored by the PATROL Agent.
%{NEWLINE}, %{\n} Linefeed. This is useful to produce multi-line information.
%{OBJECT_CLASS} Class of the instance that triggered the alert. Example: MS_HW_FAN
%{OBJECT_DEVICEID} Hardware Sentry internal device ID of the instance that triggered the alert. Example:1.1
%{OBJECT_ID} PATROL internal ID of the instance that triggered the alert. Example: MS_HW_DellOpenManagehdf_11
%{OBJECT_LABEL} Display name of the instance that triggered the alert. Example: Fan: 1.1 (CPU1)
%{OBJECT_TYPE} Type of the device that triggered the alert. Example: Fan
%{PARAMETER_NAME} Name of the parameter that triggered the alert. Example: PredictedFailure
%{PARENT_CLASS} Class of the object that the faulty instance is attached to. Example: MS_HW_ENCLOSURE
%{PARENT_DEVICEID} Hardware Sentry internal device ID of the object that the faulty instance is attached to. Example: 1
%{PARENT_ID} PATROL internal ID of the object that the faulty instance is attached to.
%{PARENT_LABEL} Display name of the object that the faulty instance is attached to. Example: Computer: Dell PowerEdge 1600SC
%{PARENT_TYPE} Type of the object that the faulty instance is attached to. Example: Computer enclosure
%{PROBLEM} Description of the problem encountered by the monitored device. Example: The speed of this fan is critically low (1503 rpm).
%{RAW_VALUE} Raw value of the parameter that triggered the alert. Example: 67.30000
%{RECOMMENDED_ACTION} Recommended action to solve the problem. Example: Check if the fan is really no more cooling the system. If so, replace the fan.
%{SYSTEM_DOMAIN} Name of the domain the monitored system belongs to.
%{SYSTEM_FQDN} FQDN (fully qualified domain name) of the monitored system.
%{SYSTEM_HOSTNAME} IP address or full name as specified by the user when adding the new monitored system (i.e. MS_HW_MAIN/<id>/hostname).
%{SYSTEM_IP} IP address of the monitored system.
%{SYSTEM_METAFQDN} MetaFQDN of the monitored system (i.e. FQDN/IPAddress).
%{SYSTEM_NAME} Name of the monitored host or hostname specified while configuring the server monitoring.
%{SYSTEM_TOKENID} TrueSight OM device Token-ID (i.e. MS_HW_MAIN/<id>/MetaTokenID).
%{TIME} Current time in the HH:MM:SS format. Example: 11:14:53. The available formats for the %{TIME:…} macro are listed in the Format Symbols for %{ASCTIME:…} macros section of this documentation.
%{VALUE} Formatted value (with unit) of the parameter that triggered the alert. Example: 67.3 °C

Alert Delay

To prevent false alerts, you can set the number of times thresholds can be breached before triggering an alert on the following parameters:

  • Numeric Parameters (Temperature, Speed, Error Count, etc.)
  • Discrete Parameters (Predicted Failure, Need Cleaning, Status, etc.)
  • Connector Status Parameters (for MS_HW_CONNECTOR class only)
  • Present Parameters (when a device is reported as missing)

KM Engine

Configuration Mode

Hardware Sentry can either be used with the Monitoring Studio Web console (Classic Mode) or BMC TrueSight Operations Management (CMA Mode). When the KM is used with BMC TrueSight, operation settings cannot be customized via the Web interface. To force the configuration from the Web console, turn ON the Force Classic Configuration Mode.

You will be prompted to reinitialize Hardware Sentry. Click YES to reinitialize the KM immediately or NO to save this setting in the configuration. When you are ready to reinitialize Hardware Sentry, or if you want to customize the reinitialize options, use the REINITIALIZE button at the bottom of the page. Refer to the Reinitialize section for details.

Once forced to Classic Mode, the KM stops processing the Hardware Sentry policies from BMC TrueSight.

Error Count Collection

Hardware Sentry automatically clears alerts on the Collection Error Count parameter after 2 hours and 15 minutes. You can customize this property by entering another value in Automatically Clear Alerts on the KM’s Collection Error Count After….

Hosts Attachment

By default, all monitored systems are created under the Hardware Sentry icon in the Web console. However, the display in your monitoring environment can be customized to help you organize your monitoring environment and facilitate the identification of your monitored systems.

Property Description Default
Place Hosts under the Hardware Sentry KM Icon To display all hosts or groups of hosts directly under the Hardware Sentry icon. If turned OFF, hosts or groups of hosts will automatically be displayed under the KMs icon (PATROL Agent). ON
Prefix Hostnames with To add a string that will prefix all hostnames in the Web console. Leave empty to display the hosts or groups of hosts with their respective names. Hardware on
Gather Hosts Belonging to a Group under a Container To group the hosts in a container (MS_HW_HOSTGROUP class). If turned OFF, hosts will be displayed under the Hardware Sentry icon ON

Tuning

Hardware Sentry allows you to customize several configuration properties that can impact the KM performance.

Property Description Default
Discovery Interval The frequency at which Hardware Sentry performs discoveries to detect new hardware components or detect those that have gone missing. 1 hour
Polling Interval The frequency at which Hardware Sentry polls the managed systems to collect hardware health data. 2 minutes
Maximum Number of Discovery Threads The maximum number of concurrent discovery threads allowed. 10 thread(s)
Maximum Number of Concurrent Collect Threads Per Host The maximum number of concurrent collect threads allowed per host. 15 thread(s)
Maximum Number of Concurrent Remote Connections Per Host The maximum number of concurrent remote connections allowed per host. 5 connection(s)

Parameter Alert Settings

Hardware Sentry enables you to customize the alert behavior for some critical parameters:

Parameter Description Default
Chassis Intrusion Detection To trigger alerts if the chassis of your managed system has been opened. An unexpectedly opened chassis could imply that the system is physically accessed by an unauthorized person. Upon Opening or Breach of a Closed Chassis
Unknown Status Value Interpretation To configure Hardware Sentry to perform specific actions when the KM finds an unexpected value for a monitored device and returns the poll with “Unknown Status”. Set the Status to OK (0)
Automatically Clear Alerts on Error Count Parameters After… To automatically acknowledge an alert on ErrorCount parameters and reset them to zero after a specified period of time. 1 hour
Missing Device Detection To enable (ON) or disable (OFF) the detection of missing devices. When a device that was previously detected in the system is no longer found, its Present parameter goes into alarm.

The missing device detection feature does not apply to logical disks, voltage and temperature sensors, LEDs, and LUNs.
ON
Remove Missing Devices To specify when Hardware Sentry should remove missing devices from the Web console.

The Missing Device Detection must be enabled (ON).
Never
Clean-up Missing Devices on Startup To remove all missing devices from the Web console upon a PATROL Agent restart. OFF

Advanced Settings

Internal Message Notification

Hardware Sentry can be configured to inform you of any internal problem that may occur or any special operation that may be performed. The internal message notifications can help resolve any potential issue, other than hardware problems – which are reported through the regular system of warnings and alerts.

Property Description Default
Write Internal Messages to the PATROL System Output Window To select the messages you would like to be notified about (in System Output Window and/or PATROL Events). All
Trigger PATROL Events for Internal Messages To select the type of PATROL Event the messages will trigger. None
Unknown Status Value Trigger an Internal KM Issue Notification To enable (ON) or disable (OFF) internal notifications when an unknown status is collected. ON

Threshold Management

By default, Hardware Sentry automatically determines which mechanism (Tuning or Event Management) is best suited to the managed system when it first runs. This threshold mechanism mode can however be modified by setting the Threshold Management Mode to:

  • Tuning: Hardware Sentry manages its thresholds through the standard internal PATROL mechanism (override parameters). Thresholds are stored in the PATROL Agent configuration under the /tuning tree.
  • Event Management: the KM manages its thresholds through the Event Management mechanism. Thresholds are stored in the PATROL Agent configuration under the /AS tree. This option requires that you set up PATROL for Event Management KM on your PATROL Agent. PATROL for Event Management has to be enabled and preloaded.
  • No Thresholds: Hardware Sentry sets no thresholds on the monitored objects. You are required to set them manually.

Timeouts

Hardware Sentry is configured to use optimal timeouts for communication protocols and processes. Timeouts may have significant impact on the performance of your monitoring environment. We strongly recommend that you consult with Sentry Software Support before modifying these settings.

Deactivate Device Class

Hardware Sentry allows you to set aside device classes from the monitoring process and, therefore, prevent the KM to collect metrics and performance data for these specific devices. By default, this option is disabled (OFF).

To deactivate specific device classes:

  1. Turn ON the Deactivate Device Classes option to display the list of devices classes
  2. Turn ON each device class you wish to deactivate.
  3. Click SAVE

Download.CFG

You can export Hardware Sentry global settings as a pconfig file (.CFG) that can be imported into a TrueSight CMA policy. All the Agents using the policy will therefore share the same Hardware Sentry Global Settings.

Java Settings

To operate properly, Hardware Sentry requires that Java 1.8.00 or higher and a Java Runtime Environment (JRE) are installed on the same system that runs the PATROL Agent.

The Java Settings page displays information about the JRE currently in use and a list of all the JREs installed on the monitored system. You can also configure the following options to fine-tune the Java properties:

Options Description
Automatically Select Appropriate JRE To let Hardware Sentry automatically find and use the most compatible JRE upon the initial discovery that occurs when the PATROL Agent and Hardware Sentry start.
Execute the JVM as an Alternate User By default, the Java process is launched with the same credentials as the PATROL Agent. If the PATROL Agent’s default account does not have sufficient privileges to perform the operations required by the Java Collection Hub, another user account can be provided.
Java Minimum Heap Size To customize the minimum size (in MB) of the Java Hub Heap Memory according to your environment requirements. By default, the minimum size is set to 128 MB.
Java Maximum Heap Size To customize the maximum size (in MB) of the Java Hub Heap Memory according to your environment requirements. By default, the maximum size is set to 512 MB.
Disable JRE Pre-Validation By default, Hardware Sentry checks that the selected JRE meets the requirements. However, you can disable this option to bypass the compatibility check. Use this option only if you are certain that the provided JRE is compatible even if the compatibility check fails. An unsuitable JRE will prevent Hardware Sentry to operate properly.
JVM Arguments To provide arbitrary arguments to the java -jar … command line that Hardware Sentry uses to launch the Java Collection Hub in the Command line options field.

The JVM needs to be restarted for your new settings to be taken in account.

Keywords:
hardware km patrol