lpar2rrd

lpar2rrd is a free, open source and web-based application that allows you to generate historical, future trends, real-time CPU usage graphs of LPARs (logical partitions) and shared CPU utilization of IBM Power servers. It has been engineered to support only HMC-based micro-partitioned systems that have a shared CPU pool.

The software can also be used to collect complete logical and physical (HW) configuration of all LPARs, managed systems, and all changes in their configuration and state. It supports all sorts of logical partitions, including VIOS, AS400, Linux and AIX, and its agentless, which means that it gets all the information from IVM, SDMC or HMC.

Features at a glance

Key features include support for a wide variety of LPARs and operating systems, support for creating charts based on collected data, support for automatic creation of a menu-based web interface for viewing charts, logs and configuration, support for accessing HMC servers via SSH keys, support for displaying the total memory usage for each managed system.

In addition, the application displays the last 100 changes that took place in the configuration, as well as the last 100 change in the state of all managed systems and their logical partitions. It is easy to use, easy to install and easy to configure, and doesn’t require additional management for adding, removing, renaming or creating LPARs.

By default, the software creates graphs with one year of historical utilization data, but it can be easily configured to generate graphs with more than one year of historical CPU usage. It automatically loads all the historical data that is collected on the HMC and saves it in RRDTool databases.

Under the hood and availability

The program is written in the UNIX Shell and Perl programming languages, distributed as a single, universal sources archive, designed from the ground up to be installed on any distribution of Linux, supporting both 32-bit and 64-bit instruction set architectures.

What is new in this release:

• VMware support

What is new in version 4.80:

• VMware support

What is new in version 4.70:

• GUI enhancements
• Support of new POWER8 high-end models in CPU Workload Estimator
• NMON online grapher
• New main configuration summary page

What is new in version 4.60:

• GUI enhancements
• Support of new POWER8 high-end models in CPU Workload Estimator
• NMON online grapher
• New main configuration summary page

What is new in version 3.50:

• A CPU Configuration Advisor was added. It aims to provide automatic verification of CPU logical setup for all your LPARs and POOLs based on historical utilization data.

What is new in version 3.40:

• This version supports lpar renaming and much faster processing.

What is new in version 3.32:

• This version fixes two issues which appeared in 3.30.
• Check the lpar2rrd 3.30 release notes for a manual fix of 3.30 if you already running that version.

What is new in version 3.30:

• This version enhances the CPU Workload Estimator and Active Memory Sharing support.

What is new in version 3.15:

• Changes include a CPU workload estimator.

What is new in version 3.11:

• This release contain fixes and enhances for "Physical and Logical cfg".

What is new in version 2.61:

• is no longer only off-line tool, since 2.59 you can do whenever refresh (via a WEB link) to get nearly "real-time" graphs
• intended only for micro-partitioned systems with shared CPU pool
• creates charts based on utilization data collected on HMC's/IVM's (lslparutil)
• agent-less, no agents need to be installed on LPAR's
• ssh-keys based access to HMC/IVM servers to get all the data, so it does not cause ANY load on monitored LPAR's
• support of all types of LPAR's and OS'es : AIX , VIOS , Linux on Power, i5/OS on IBM Power Systems (formerly pSeries/iSeries)
• automatically created a menu based WWW front-end for viewing charts, configuration and logs
• physical (HW) and logical configuration inventory of all managed systems and their lpars (once a day)
• showing 100 last changes in configuration and 100 last changes in the state of all managed systems and their lpars
• showing total memory usage for each managed systems
• LAN income/outcome traffic of IVE/HEA adapters and error packets
• simple to install, configure and use (initial install & configuration together with supporting tools like Apache/Perl/SSH should not take more than half an hour!)
• no any additional management when any change/adding/remove/rename of LPAR or managed system (the tool discovers everything automatically for all configured HMC's/IVM's)
• default graphs are created 1 year back if historical utilization data on HMC/IVM is available (note the HMC keeps 60sec sample rate data for 2 days, hourly data for last 2 months and daily data for last 2 years if utilization data collection is enabled)
• for viewing older data than 1 year you have to use Historical reports if the data is already present in LPAR2RRD
• initially the tool loads all historical data already collected on the HMC/IVM, after that the tool loads only new data every hour (depends on cron schedule), saves it in RRDTool databases and re-draws the graphs
• it can be hosted on any *NIX (theoretically even WIN) platform, it just needs a web server, SSH, Perl and RRDTool installed. (check prerequisites below)
• by default is created 4 kind of graphs for each LPAR, CPU shared pool and memory usage. First 3 (last day, week and month) are based on 60secs averages, yearly chart is based on daily averages, all provided by the HMC

What is new in version 2.59:

• Added "real-time" refresh of lpar/pool graphs for HMC utilization sample rates < 1 hour, however the refresh is not automatic, but manual (a "refresh" link in the web).