performance – Eclipse Systems Knowledge Base

How do I install nmon on Linux?

To download and install nmon on your Linux server:

mkdir -p /u2/UTILS/bin && cd /u2/UTILS/bin
wget http://sourceforge.net/projects/nmon/files/nmon_x86_rhel52
chmod +x nmon_x86_rhel52
ln -s nmon_x86_rhel52 nmon

How do I monitor system performance using nmon?

nmon is a free console-based performance monitoring tool from IBM. It works on both AIX and RHEL.

If nmon is not already installed, you must first install it:

How do I install nmon on Linux?
How do I install nmon on AIX?

To use nmon:

View a step-by-step screencast of this process:

Log into your server as root using a terminal emulator such as PuTTy (see also: How do I use PuTTy with my Eclipse server?)
(AIX only) Set the terminal emulation type to match the PuTTy console:
```
export TERM=xterm
```
Run:
```
nmon
```
Press the appropriate keys to launch or close various monitoring windows. Type “h” for help. Here are some common views:
- “t”: display top processes
- “t” then”u”: display top processes by CPU utilization and show command line arguments
- “c”: display CPU utilization
- “l”: display longer-term CPU utilization averages
- “d”: display disk I/O graphs
- “h”: display help

Resources:

IBM DeveloperWorks Wiki: nmon for AIX Performance Monitoring
IBM DeveloperWorks: nmon performance: A free tool to analyze AIX and Linux performance
nmon

What should I monitor on my Eclipse servers?

The basic elements that make up monitoring are:

Events: triggered when a set condition occurs
Threshold: the point on a scale that must be reached in order to trigger a response to an event. The response can be an alert, a notification, or a script being run.
Notification: how an administrator is informed that something (an event or a response) has occurred.
Health: a set of metrics that define the state of the functionality being monitored. The administrator defines the values that represent a “healthy” state for each of their components.

For your Eclipse environment, we recommend monitoring the following items where applicable and possible:

General Monitoring:

Availability (ping test)
CPU (high CPU threshold alerts, tracking historical trends)
Memory (high RAM threshold alerts, tracking historical trends)
Disk space (top utilization threshold alerts, tracking historical trends)
Hardware (failures, power loss, firmware events, etc.)

Database server:

UniVerse (process running, responsive, spooler status)
SYSTEM.ADMIN (process running)
SOCKET.PH.SERVER (process running, listening on port 22222)
VSIFAX (process running, responding, modems down)
JBoss (process running, listening on port 2080)
Samba shares (availability, read/write)
CUPS (print queue status)
Sendmail (service running, queue status)

Forms Server:

Windows share(s) (availability, read/write)
Formscape (availability of port or services running)
VSIFAX (services)

Imaging Server:

Windows share(s) (availability, read/write)

Internet Gateway Server:

IIS (service running, listening on ports 80 and/or 443)