In this post, we'll use Ops Center to add hardware monitoring to Enterprise Manager. We'll discuss the existing capabilities of Host targets, show how to create an Infrastructure Stack and demonstrate some of the features it provides to Enterprise Manager.
A recording of this community call is now available here:
This blog post uses both the Enterprise Manager 12c Cloud Control and Ops Center products. The following list describes the initial setup state and provides links to Oracle documents you can use to install and configure both products:
- Enterprise Manager 12c
- Install and configure an Oracle Management Server and Repository (OMS and OMR)
- Install the Oracle Management Agent (OMA) on a target OS instance
- Ops Center 12c
- Install and Configure the Enterprise and Proxy Controller (EC and PC)
- Discover and manage the system and its associated OS (installing the Ops Center agent on the OS instance)
I've configured the environment for this post in the following way:
- Enterprise Manager 12c: OMS and OMR running separately
- Ops Center 12c: EC and PC running co-located
- Two sample systems, both running Solaris 11
- An Oracle SPARC T4-2 server
- An Oracle SunFire X4200 M2 server
By default, installing an Enterprise Manager agent on an OS instance creates an associated Host target. A Host provides a lot of useful data about the platform that hosts the OMA, such as
- CPU and memory utilization
- File system size and utilization
- Network interfaces and activity
- Program and process resources
- User activity
The Host does not provide sensor data associated with the server and cannot report issues with the underlying hardware. Fortunately Ops Center can do both of these things, and you can incorporate server data, monitoring thresholds and alerts into your Enterprise Manager environment.
To connect Ops Center with Enterprise Manager, select the left-hand Navigation link Administration > Enterprise Manager Cloud Control. Click on the right-hand Action link Configure/Connect and open a pop-up dialog box. The dialog lets you configure the OMS and OMR settings. The screen shot below shows both steps from the wizard.
Enterprise Manager 12c provides a deployable plug-in to manage an Infrastructure Stack. For Enterprise Manager installations running in online mode, you can download it from the Extensibility > Plug-Ins menu from the Servers, Storage and Network category.
Download and deploy the plug-in to the OMS. You can immediately deploy it to the OMA instances you want to integrate with Ops Center, or wait until you create an infrastructure stack. (Enterprise Manager will automatically install the plug-in to the OMA if it is not already present)
An Infrastructure Stack associates data from a system in Ops Center with targets in Enterprise Manager. To create one, select the menu option Setup > Add Targets > Add Targets Manually. From the wizard, select Infrastructure Stack from the pull-down menu and identify the Monitoring Agent that will be used.
The subsequent configuration screen allows you to define the name
for the target, to identify the Enterprise Controller that will provide
the data for the server, and to specify the Ops Center login
credentials. Any user account defined in Ops Center is suitable for the target.
What benefits does an Infrastructure Stack provide? As the consolidation point for server-related information, it enables you to perform three principal tasks:
- Monitoring, with metrics and thresholds for the server
- Reporting, using a set of standard Information Publisher reports
- Incident Management, with Ops Center hardware alerts
Let's look at some examples for each.
An Infrastructure Stack provides a wealth of information about a server, including identification information, state, capabilities and sensor data. The home page provides a summary of current values for power consumption, temperature, fan speed, and reported incidents.
The metrics section provides more detailed data such as sensor values, thresholds, installed firmware and server capabilities.
Note that the reported values ultimately depend on what data is available for a specific type of server. Some earlier models don't report temperature data, for instance.
Enterprise Manager provides three standard Information Publisher reports:
- Infrastructure Stack Topology, showing the server and OS associated with the stack
- Infrastructure Stack Configuration, providing a tabular summary of key data
- Hardware Sensors, showing current values and thresholds for monitored server data
The following screen shots provide sample report output for the infrastructure stack configuration and hardware sensor data.
Incident reporting is an optional capability for an Infrastructure Stack. Enabling the feature causes Ops Center to forward hardware alarms, allowing you to consolidate problem management in Enterprise Manager.To enable incident reporting, navigate to the Infrastructure Stack and select the menu option Monitoring > Metric and Collection Settings. Select the link Collection Schedule for Infrastructure Stack Alarms to edit the settings:
If you toggle the collection schedule to Enabled, Enterprise Manager will activate incident reporting based on hardware alarms. By default, the data refresh frequency is once every five minutes, with warning or critical alarms being reported.
In this example, we simulated an hardware alarm using the IPMItool utility from the Hardware Management Pack. Ops Center forwarded the event and all associated data, which generated an
actionable incident in Enterprise Manager,
In this post, we have demonstrated how to integrate Ops Center data into Enterprise Manager, and described the features available for an Infrastructure Stack. If you would like to learn more, please join us for the WebEx demo on May 9th.