Oracle® Enterprise Manager Cloud Administration Guide 12c Release 2 (12.1.0.2) Part Number E28814-03 |
|
|
PDF · Mobi · ePub |
This chapter covers usage of the Consolidation Planner feature provided with Oracle Enterprise Manager Cloud Control.
This chapter includes the following sections:
Over the years, the typical enterprise data center will grow steadily due to the addition of servers required to satisfy the needs of the business. This growth typically results in excess servers that occupy rack space, consume a lot of power for cooling, and require system maintenance such as security and patching.
Depending on the procurement cycle or the specific hardware vendor agreement in effect, enterprises may acquire different types of server hardware and operating systems, inadvertently creating a confusing array of systems that administrators need to manage, administer, patch and upgrade. This in turn increases labor and ongoing maintenance and support costs against IT budgets. Enterprises look at consolidation as a way to migrate their disparate systems onto standardized operating systems and hardware, with the ultimate goal of reducing costs.
Enterprises also are increasingly investigating virtualization technologies such as Oracle Virtual Machine by moving from physical to virtual servers. This makes it possible to use the shared hardware infrastructure while getting the benefits of isolation that virtualization provides.
The goal of consolidation is to identify such under-utilized servers and find a way to consolidate them, enabling the enterprise to free up as many servers as possible while continuing to maintain service levels. Since servers have different levels of CPU capacity, Consolidation Planner uses computer benchmark data to normalize the CPU usage for a given hardware in the consolidation process. Specifically, Consolidation Planner uses the SPECint©_rate_base2006 benchmark, published by Standard Performance Evaluation Corporation (SPEC), for the different classes of hardware.
The Consolidation Planner feature enables you to match managed servers you want to consolidate with the generic physical machines, Oracle Exadata database machines, or Oracle Virtual Machine (VM) servers they can be consolidated to. By leveraging metric and configuration data collected from managed target servers by Enterprise Manager Cloud Control, Consolidation Planner helps you determine the optimum consolidation scenarios that also reflect business and technical constraints in the consolidation process.
The following concepts are central to using Consolidation Planner.
An existing server that will be considered for consolidation.
An existing or yet-to-be-purchased server that a source server will be consolidated to. Can also be thought of as the consolidation target. These may be individual machines or an integrated system such as the Oracle Exadata Database Machine.
Defines the scope of a potential consolidation effort, including:
The type of consolidation. In the current release, two types of consolidation schemes are supported:
P2V: From physical source servers to Oracle Virtual Machine (VM) destination servers
P2P: From physical source servers to physical destination servers
The preliminary set of candidate source servers to consider consolidating from
The preliminary set of candidate destination servers to consider consolidating to
The duration over which data used to generate consolidation scenarios will be collected for the source servers
Each consolidation project contains one or more consolidation scenarios that are generated based on the inputs provided. Inputs provided to a scenario include:
The source server resource requirements that a destination server must meet, including one or more of the following: CPU, memory, disk I/O, network I/O, and disk storage
Any business, compliance or technical constraints that must be considered
The destination servers to consider in the scenario
A set of pre-configured consolidation scenarios are provided, representing conservative, aggressive, and medium consolidation schemes. Each scenario is generated based on inputs you provide. Alternatively, you can create your own custom scenarios that best suit your situation. Once created, you can compare the various scenarios to determine which consolidation strategy best meets your requirements.
Each scenario also includes initial mappings between each source server and the destination server(s) it may be consolidated to. You can choose to create mappings manually, or allow Consolidation Planner to create them automatically. Once all inputs are specified, you can run the scenario and evaluate the results. You can change the values in a scenario, then re-run it to view the new results.
Consolidation Planner allows you to specify various constraints that must be considered when creating consolidation projects and scenarios.
This section covers the following:
For source servers, you define constraints on the basis of compatibility or exclusivity.
Servers are considered compatible if they match on certain property and configuration values.
Properties include the following:
Lifecycle Status
Department
Location
For example, there may be situations that mandate that specific servers remain within a specific location, such as data center location or geography.
Configuration items include the following:
Network Domain
System Vendor
System Configuration
CPU Vendor
CPU Name
Operating System
For example, there may be an accounting policy that the same system and CPU vendor be used.
You can choose to establish compatibility on none, some, or all of these target properties and configuration items.
You can choose to exclude servers from consolidation scope because they violate certain Oracle best practices. Set either or both of the following conditions to exclude matching servers:
Nodes of a RAC Database
Nodes of an Oracle Cluster
For destination servers, you can scope candidates to either new or existing candidates, but constraints are primarily expressed as a percentage of CPU and memory resource utilization; that is, how much of either resource type can maximally be used on a destination server.
The steps in the consolidation planning process are:
Create a consolidation project. See Creating a Consolidation Project.
Define one or more consolidation scenarios within the project. You have two options:
Use a pre-defined consolidation scenario. See Using a Pre-configured Consolidation Scenario.
Create a custom consolidation scenario. See Creating a Custom Consolidation Scenario.
Evaluate your consolidation scenarios in the Consolidation Planner console to determine the consolidation strategy that best meets your needs. See Evaluating Consolidation Scenarios.
Modify the settings of your scenarios to generate different results. Continue this process until you have the most optimal scenario(s) for your situation.
Note that in this release of Enterprise Manager Cloud Control, consolidation scenarios are created for planning purposes only. Execution of scenarios - that is, the actual movement of software or data from source servers to destination servers - is not supported.
When consolidating multiple source servers to more than one destination server, the resource requirements of source servers are checked against the resource capacity of the destination servers. To consolidate all identified source servers to the least number of target servers, Consolidation Planner tries to identify a set of source servers that have known resource requirements that will fit into a destination server's corresponding available resources as tightly as possible.
For example, if the available memory in a target server is 2 GB, Consolidation Planner will try to find a set of source candidate servers with sum of required memory as close to 2 GB as possible then “fit” the source servers in the target server. The goal is to “fit” the source servers into the least number of destination servers.
This section includes the following:
You will create a consolidation project for each consolidation effort, then add individual consolidation scenarios within it. You can then compare consolidation scenarios to determine which consolidation strategy makes the most sense.
After the project is defined, a Cloud Control job is submitted to collect available data for the specified servers from the Management Repository. Once the job finishes, the project becomes an active project. As long as the project is in an active state, data collection will continue.
From the Enterprise menu, select Consolidation Planner.
Click Create Project.
Enter the consolidation project name.
Select the consolidation type. Two types of consolidation schemes are supported: From physical source servers to Oracle VM servers (P2V), and from physical source servers to physical servers (P2P).
Note that for both types, only source servers running on an operating system listed in the pull down menu can be consolidated.
Select the source servers to be added.
Click Add Source Servers to see a list of managed servers that could potentially be consolidated. Select the servers you want to add, then click Select.
Optionally click Remove to remove a server from the list.
Optionally select one or more existing servers to consolidate the source servers to.
If you are consolidating from physical servers to Oracle Virtual Servers (P2V), click Add Existing Virtual Servers to view a list of existing Oracle Virtual Machine destination servers to consolidate the source servers to. Select the servers you want to add, then click Select.
If you are consolidating to an Oracle Virtual Machine server pool, the first virtual server pool is selected by default to provide target server candidates.
If you are consolidating from physical servers to physical servers (P2P), click Add Existing Database Machines to view a list of Oracle Exadata Database Machine servers to consolidate the source servers to. Select the servers you want to add, then click Select.
Specify the duration over which data used to generate consolidation scenarios will be collected for the source servers specified in the project in the Data Collection region. This data will be used to determine the resource requirements that a destination server must meet.
Specify the minimum number of days the collect data. The default minimum value is 21 days. To use existing historical data to run and view consolidation scenarios immediately, set the minimum number of days to 0.
Specify the maximum number of days the collect data. The default maximum value is 90 days.
Specify when to begin the data collection process.
Optionally select Continue Data Collection Over the Maximum Days to purge the oldest day's data when data for a new day is added.
Click Pre-configured Scenarios if you want to use one or more of the out-of-the-box consolidation scenarios.
The pre-configured scenarios will be generated when the project is created using data collected for the source servers defined in the consolidation project.
You can also opt to create your own custom scenario once the consolidation project has been completed.
Click Submit when finished.
Once the project is created, it will show up in the Consolidation Planner Console. Consolidation scenarios can then be defined for this project.
When creating a consolidation project, you can optionally choose to generate three pre-configured consolidation scenarios to add to the project. These out-of-the-box scenarios represent conservative, aggressive, and medium consolidation schemes.
These scenarios are generated using data collected for the source servers defined in the consolidation project at the time the project is created. If no data is available when the project is created, the pre-configured scenarios will be automatically generated once the minimum amount of data has been collected.
During consolidation project creation, click Pre-configured Scenarios.
Select the scenario(s) you want to add to the project. Note that you can modify the name of each scenario.
By default, the scenarios are designated by the method used to aggregate daily source server resource usage:
Aggressive: Aggregate data based on average daily usage per hour.
This typically results in a high consolidation (source server:destination server) ratio, because more source servers will “fit” into each destination server. But because more source servers are involved, the odds that one or more will not meet the resource requirements are higher.
Conservative: Aggregate data based on maximum daily usage per hour.
This typically results in a lower source server:destination server ratio, because fewer source servers will “fit” into each destination server.
Medium: Aggregate data based on the 80 percentile usage.
This typically results in a source server:destination server ration somewhere between Aggressive and Conservative aggregations.
Usage statistics are calculated based on the following criteria:
Resource Requirements: The source server requirements, such as CPU, memory (GB) and disk capacity, that must be met by destination servers.
Applicable Dates: The days of the week on which resource usage metrics are collected. Typically you will specify the day of highest resource usage.
Target Server Utilization Limit: The maximum resource utilization (percentage) that can be used on destination servers.
Click Use New Servers if you plan to use destination servers that do not exist yet, but you expect to be provisioned or purchased (also known as phantom servers).
Fill in the data for each server. For CPU capacity, provide the estimated SPEC© rate if available; otherwise, the Consolidation Planner will determine requirements based on the CPU configurations provided earlier in the wizard.
The Consolidation Planner will determine how many destination servers are required as part of the consolidation scenario generation.
Click Use Existing Servers to use the destination candidates specified in the Consolidation Scope section of the consolidation project.
Note that you can edit the CPU (SPEC© rate) and disk I/O fields.
Click OK when finished.
You can create custom consolidation scenarios instead of or in addition to using the pre-configured scenarios. Multiple scenarios can be created within a project, enabling you to compare different scenarios before deciding on a solution. New consolidation scenarios are created within an existing consolidation project.
A wizard guides you through the creation process.
Note that you can also opt to use the pre-defined consolidation scenarios provided.
From the Enterprise menu, select Consolidation Planner.
Click on the consolidation project you will create the scenario within.
Click the Create Scenario button.
Specify the scenario details, such as scenario name.
Specify the source server resources to consider. The Consolidation Planner will aggregate the specified resources to determine the total requirements.
Resource Type: The server requirements, such as CPU and memory (GB), that must be considered.
Days: The days of the week on which resource usage metrics are collected.
Resource Allocation: The method used to aggregate daily source server resource usage. Values are:
Aggressive: Aggregate data based on average daily usage per hour.
This typically results in a high consolidation (source server:destination server) ratio, because more source servers will “fit” into each destination server. But because more source servers are involved, the odds that one or more will not meet the resource requirements are higher.
Conservative: Aggregate data based on maximum daily usage per hour.
This typically results in a lower source server:destination server ratio, because fewer source servers will “fit” into each destination server.
Medium: Aggregate data based on the 80 percentile usage. This typically results in a source server:destination server ration somewhere between Aggressive and Conservative aggregations.
The date ranges should define a period of time that is typical of standard resource requirements.
Click Estimated Requirements to view the estimated total resource requirements.
Resource requirements are shown based on the averaged hourly requirement. The 24-hour requirement pattern will be used as the minimum requirements that must be met by consolidation target(s).
Click Next to define consolidation constraints.
Specify business, corporate or technical constraints that must be enforced. These constraints will be used to:
Guide the server allocation process during automatic source-to-destination server mapping; or
Calculate violations if manual mapping between source and destination servers is used
Specify compatible servers.
Servers are considered compatible if they have the same specified target properties and configurations. If you are consolidating multiple source servers to a single target server, only compatible servers can be consolidated together on the same target server.
Specify mutually exclusive servers.
Certain types of source servers are mutually exclusive and should not be consolidated together on the same destination server due to various reasons. For example, database nodes in an Oracle RAC cluster should not be consolidated on the same destination server. Contractual obligations may also restrict certain applications from running on the same server.
Optionally click Preview Effect of Constraints to view a list of source servers that are not compatible based on the defined constraints.
Click Next to specify destination server candidates.
Click Use New Servers if you plan to use target servers that have yet to be provisioned or purchased. The Targets Planning region shows the minimum required target servers based on the resource requirements of all servers in the scenario.
If you are defining the scenario for a P2V project, fill in the data for each server. For CPU capacity, provide the estimated SPEC© rate if available; otherwise, click the List of Values icon next to the CPU capacity input field, then select a server configuration that most closely matches your needs.
If you are defining the scenario for a P2P project, in addition to the same new server option as a P2V project, you can also select a pre-configured Oracle Exadata database machine as the planned target server. Click the List of Values icon next to the read-only Configuration input field, then select one of the pre-configured Oracle Exadata database machines.
The Consolidation Planner will determine how many target servers are required as part of the consolidation results.
Click Use Existing Servers to specify a set of existing managed servers to use as targets.
These are the servers you specified when defining the scope for the consolidation project. The Consolidation Planner will determine the available hardware resources based on collected usage data.
By default, the consolidation process will try to use as few target servers as possible. If you prefer, you can choose to balance the source load across all destinations.
Enter percentages for Maximum Allowed Resource Utilization on Destination Servers. The defaults are 75 for CPU and 90 for memory, respectively.
Click Next to map the source servers to the destination servers.
Next, map the source servers to the destination servers they will be consolidated to. The objective is to fit source server requirements with each destination server's available resources as tightly as possible.
It is recommended that you allow Consolidation Planner to perform this mapping automatically. This will allow the tool to maximize resource utilization of destination servers based on server resource capabilities and the various consolidation constraints specified.
If you prefer to map each source and destination server manually:
Click a source server in the list.
Click the “flashlight” icon to select the destination server to map to the source server. Note that you can map a single source server to a target server or multiple source servers to a target server, but there can be only one target server.
The resulting consolidation report will show any resource and/or constraint violations due to such manual mapping.
Click Next to review the consolidation scenario.
Click Submit to save the new scenario.
You can view details for your consolidation scenarios using the Consolidation Planner console. After evaluating the consolidation scenario results, you can define different plans as well as re-run existing scenarios with new inputs, then compare the results again. This iterative process helps you obtain the optimized consolidation scenario which is generated by compromising various factors and weighing different trade-offs.
Compare the consolidation scenarios you create to determine which consolidation strategy best meets your requirements.
Your objective is to:
Match source server resource requirements with destination servers best able to meet those requirements
Fit source server requirements with each destination server's available resources as tightly as possible, so you can get maximum usage of destination server capacity
Optionally balance the source server workload across all available destination servers
Note that in the current release, consolidation scenarios can be created for planning purposes only. Consolidation scenarios cannot be executed using Consolidation Planner.
From the Enterprise menu, select Consolidation Planner.
First, examine the project containing the scenario you want to view.
The Status column indicates the status of data collection, based on the minimum and maximum collection days specified for the project.
Click the Source Servers tab to view CPU, memory and disk storage data and utilization rates collected for the source servers defined in the project.
Click the Source Workload tab to view a graph showing source server resource usage data collected. Data is shown for each 24 hour period.
Click the Report button above the table when the project is selected to view summarized information and more details.
Next, view the data for a specific scenario. Click on any metric or on a tab at the bottom of the page to view additional details. Key data includes:
Destinations: The list of destination servers to which the source servers will be consolidated. Resource configuration and calculated utilization are shown for each destination server.
Ratio: The ratio of source servers to destination servers. By default, the Consolidation Planner will try to “fit” source servers into as few destination servers as possible.
Confidence: The percentage of the data collected for source servers that meet the source server usage requirements defined in the scenario. This value is aggregated for all source servers defined with the project.
Violations: The number of violations of technical or business constraints defined in the scenario.
Exclusions: The number of source servers that do not have a qualified mapping to a destination server. These are source servers that exceed the capacity of available destination servers. This metric is applicable only if auto-mapping of source servers to destination servers is used.
A different set of constraints may result in a different optimal scenario. Modify the constraints to come up with different scenario results.
Select a report in the table and click Publish Report to make report contents public. This action integrates with BI Publisher, where you can:
Save reports in a variety of formats (Excel, PowerPoint, HTML, PDF)
Distribute generated reports to e-mail lists (users who do not have access to Enterprise Manager, for example) on a defined schedule