SA Legacy – Equivalence

Requirement

Overview

A baseboard management controller (BMC) is a specialized service processor that monitors the physical state of a computer, network server or other hardware device using sensors and communicating with the system administrator through an independent connection. The BMC is part of the Intelligent Platform Management Interface (IPMI) and is usually contained in the motherboard or main circuit board of the device to be monitored. The sensors of a BMC measure internal physical variables such as temperature, humidity, power-supply voltage, fan speeds, communications parameters and operating system (OS) functions.

If any of these variables happens to stray outside specified limits, the administrator is notified. That person can then take corrective action through remote control. The monitored device can be power cycled or rebooted as necessary. In this way, a single administrator can remotely manage numerous servers and other devices simultaneously, saving on the overall operating cost of the network and helping to ensure its reliability.

IPMI defines two basic types of sensors. Threshold sensors monitor “analog” things like temperature, voltage, or fan speed. Discrete sensors monitor events or states, like entity presence, software initialization progress, or if external power is applied to the system. Both threshold and discrete sensors may generate events. Sensor events are stored in system event log (SEL). Most entries will display the SEL record id, date of event, time of event, sensor group, sensor name, and the sensor event occurrence. Some timestamps in the SEL may report a date of 1-Jan-1970. This timestamp is not necessarily incorrect. It usually indicates a hardware event that occurred before a timestamp in firmware has been initialized. For example, certain hardware components will have their internal clocks reset during a power cycle.

Design

ipmi plugin

The ipmi plugin collects information about sensors provided by BMC:

The ipmi plugin generates notifications when event is received from BMC and exposes as much information as it is provided using notifications:

Plugin configuration

The following configuration options should be supported by ipmi collectd plugin:  

Here is an example of the plugin configuration section of collectd.conf file:

  <Plugin ipmi>

    Sensor "some_sensor"

    Sensor "another_one"

    IgnoreSelected false

    NotifySensorAdd false

    NotifySensorRemove true

    NotifySensorNotPresent false

    SELEnabled false

    SELClearEvent false

   </Plugin>

Implementation details

To enable support of IPMI features OpenIPMI library will be used. OpenIPMI is callback based library and provides API to register handers. The following diagram describes the process of collecting information about the analog sensors.

Collectd IPMI plugin state diagram.

Collectd IPMI plugin callback API registration workflow

The diagram above describes the workflow of initialization and handler registration used by ipmi plugin. After OpenIPMI library is initialized, the first step is to setup the SMI (system management interface) connection to create and manage the domain. The domain update handler will be called after successful creation and ipmi plugin will register sensor entity update handlers for sensors. This will cause OpenIPMI driver to scan sensors on the system and determine the sensors belongings to entity. And invoke sensor update handler to update the list of known sensors for ipmi plugin in specified domain. Each sensor has its type and can be either threshold or discrete. IPMI plugin filters and works only with predefined types of sensors described before. If system event log (SEL) feature is enabled, threshold or discrete event handler will be registered. And when event will be received from BMC, event handler will be called and notification will be sent. According to the default OpenIPMI behaviour SEL is scanned for configurable period of time. To receive events asynchronically event handler is being registered for SMI. And when event arrives, ipmi_domain_reread_sels() is called to force domain to reread SELs.

Extended list of supported sensors

The current collectd IPMI plugin implementation supports the following types of IPMI sensors:

The plugin supports only analog sensors and selects them (creates a list of monitored sensors) based on their sensor type which looks incorrect. The IPMI sensor may be of supported type but may provide discrete value instead of analog. Thus, the plugin should select sensors based on the value that sensor provide. For this reason, the select logic has been extended (to preserve backward compatibility) to support the following sensors based on their value type:

Considerations

Configuration Considerations

Deployment Considerations

If your platform does not support BMC – this plugin will be unloaded at initialization time.

API/GUI/CLI Considerations

Equivalence Considerations

Security Considerations

Alarms, events, statistics considerations

Not all metrics will be reported as not all types of sensors are supported by IPMI plugin.

IPMI plugin registers to listen for all type of sensor events received from System Event Log (SEL).

Redundancy Considerations

Performance Considerations

Testing Consideration

The timing interval requirement needs to be taken into consideration when conducting tests.

The Tests should be carried out on a system underload as well as a relatively idle system.

Other Considerations

Impact

The following table outlines possible impact(s) the deployment of this deliverable may have on the current system.

Key Assumptions

The following assumptions apply to the scope specified in this document.

Key Exclusions

The following exclusions apply to the scope discussed in this document.

Key Dependencies

The following table outlines the key dependencies associated with this deliverable.

Issues List