Step 1: Install PlateSpin Recon
PlateSpin Recon requires the Recon client, Recon server and an SQL database. Recon ships with the open-source PostgreSQL database. However, it is recommended that you use a commercial SQL database such as Microsoft SQL Server 2000 or 2005 for maximum performance.
Installing PlateSpin Recon:
- Ensure the Recon server has enough disk space for the number of workloads you wish to monitor.
- Download PlateSpin Recon.exe.
- Double-click PlateSpin Recon.exe. The PlateSpin Recon InstallShield Wizard is displayed.
- Type a location to save PlateSpin Recon files or browse for a location by clicking the Change button. By default, the PlateSpin Recon files are extracted to a temporary folder.
- Click Install. PlateSpin Recon files are extracted and saved to the specified location.
- The PlateSpin Recon Installation Launcher is displayed.
Step 2: Discover Potential Host Servers
To inventory your physical hosts: (See Figure 1.)
Pull down the Inventory tab.
Select your inventory method. Recon includes standard inventory options, such as individual machines, virtual center, Windows domain, IP range, subnet and SNMP.
Most customers prefer to collect inventory data by IP range, subnet or CSV file.
Step 3: Monitor Your Environment
There is not much to do during the monitoring step, as PlateSpin Recon performs all the work. To monitor Windows servers, Recon pulls Perfmon data. On Linux and Solaris, the solution runs an efficient shell script to gather performance metrics.
Plan on collecting inventory and performance data for approximately 30 days. Ideally, this time frame should include a significant event, such as end-of-month or end-of-quarter workloads. (See Figure 2.)
Recon collects inventory data for all discovered systems, without requiring agents on those systems. Inventory data includes the platform on which the workload is running, operating system processes, disk drives, network adapters and installed applications. It's possible to monitor all discovered systems, but most customers prefer to monitor a subset of those systems.
Ongoing monitoring collects performance metrics for disks, memory, processors, I/O and network usage. You can drill down to review system or workload details by simply clicking on each physical or virtual server.
Step 4: Analyze Planning Reports
Having a holistic view of the operating system, system age and more is helpful for inventory and auditing. But the real value in PlateSpin Recon is the ability to create customizable reports that provide graphical views of actual performance and projected system performance.
These reports provide an accurate view of the overall usage and peak usage of each server over time, allowing you to determine the best candidate systems for virtual workloads.
To plan your consolidation and run what-if analyses, click on the Consolidation Planning module tab in the lower-left of your screen. Here you can define the candidate workloads to be used in the plan, which virtualized platform to be used and the type of hardware under consideration. (See Figure 3.)
Subscribe to Connection Magazine
Recon allows you to set your own minimum and maximum thresholds for performance-metric usage. For example, you could set maximum memory usage to be less than 50 percent.
If you are looking to purchase additional hardware, you can build sample scenarios for various hardware platforms. For example, you could build one scenario for HP servers, another one for IBM machines and another for Dell systems-each with specific configuration details.
Note: All PlateSpin Recon graphical report summaries can be easily exported to standard file formats, such as HTML, PDF, Word, CSV, Excel and images.
The Good Consolidation Candidates report determines "underutilized servers," which can be better used as virtual hosts.(See Figure 4.)
You can view both ScaleOut and ScaleUp scenarios for each platform under consideration. (See Figure 5.)
You can also create ad hoc reports for things like projected 24-hour usage profiles of each hardware platform, including specific system resources running various types of virtual machines. This allows you to project system resource bottlenecks such as CPU, memory, disk I/O or network I/O.
PlateSpin Recon also includes standard reports to help build a business case by projecting each scenario's physical server count, rack space, energy usage, cooling requirements and more.
Once you have decided on a plan, you simply export your preferred scenario into an implementation project for use with PlateSpin Migrate. (See Figure 6.)
Watch for the Next Articles
The second article in this series will show you how PlateSpin Migrate automates the implementation and testing of your consolidation plan. Look for this article in the next Novell Connection issue.
For More Information
For additional resources on consolidation planning, check out these tools:
- 5-minute online demo
- PlateSpin Recon Quick Start Guide