: i'm looking for a simple utility that will display a table of hosts,
: periodically ping these hosts, and show the results. basically a
: quick and easy way to tell what portions of the network and
: specifically which computers are up. there is a version for os/2,
: called pmping, and i'd like a program like this for unix/X.
Hmmm...if you want to monitor this on a continuous basis, you could try
NOCOL/NetConsole. The URL is:
From the README:
NOCOL/NetConsole (Network Operation Center On-Line) is a network monitoring
package that runs on Unix platforms and capable of monitoring network and
system variables such as ICMP or RPC reachability, RMON variables,
nameservers, ethernet load, port reachability, host performance, SNMP
traps, modem line usage, appletalk & novell routes/services, BGP peers, etc.
The software is extensible and new monitors can be added easily.
The software consists of a number of individual, standalone monitoring agents
that poll the various network and system parameters and put it into a common
data format. All the monitors have a common display and postprocessing
interface (such as logging, notification, etc.). The design allows running
just one set of monitoring agents and *any* number of display agents, and all
of the displays see the same consistent set of data. Additionally, each event
is assigned a severity (determined by comparing against user defined threshold
values) which is gradually escalated, thus preventing false alarms and a
customized priority notification based on the severity. There are four
severity levels ranging from Critical thru Info, and each event typically
steps through each one of these severities until it reaches its maximum
The display uses UNIX 'curses' screen management and can thus run on a large
variety of terminals. The user running the display can select the minimum
display severity- only events above this minimum severity level are displayed.
To date, the various monitoring agents developed are:
- IP ICMP monitor (using IP 'multiping')
- OSI reachability monitor (using OSI ping)
- RPC portmapper monitor (using 'rpcping')
- Ethernet load (bandwidth & pps)
- TCP port monitor
- Unix host performance (disks, memory, swap, load, nfs, collisions)
- SNMP variables monitor (RMON, Cisco router, terminal server)
- TCP data throughput monitor
- Nameserver (named)
- SNMP traps
- Usage of terminal server modem lines (busy lines)
- Appletalk route monitor (for cisco routers)
- Novell service monitor
- BGP peer status
A PERL interface is available for developing additional monitors in the PERL
programming language (it is fairly easy to add additional monitors to the
A sample NetConsole display can be viewed by logging onto 'nocol.jvnc.net'
as user 'nocol'.
Note that last paragraph -- check it out yourself and see if it's what
you want. We use it in-house, and it's met our monitoring needs and then
BTW, if you're worried about whether a host is operational (and not just
merely up), a simple ping isn't sufficient. NOCOL has rpcping, which
actually talks to the RPC portmapper on each machine, giving a much
better indication as to whether a host is "alive".
- Adrian Ho
IT Architect, Infrastructure Group
NII Division, National Computer Board