Is there any way of knowing whether another machine is down without logging
onto that machine?
For instance, I'm here on grex and want to know whether m-net is up or not.
Is there any way of doing that?

11 responses total.

This response has been erased.

I've written a network monitoring tool that you can use to monitor
remote systems.  It can e-mail or page you when a site goes down, or
any specific service on it dies.

The conventional way of doing what #0 asked about is with ping.  Here on grex
it had to be depermitted because of abuses, unfortunately.

I have it firewalled on nether.net outgoing, so nobody can do ping attacks, but
can do traceroutes, etc.. from there.

Have you (or are you going to) publish source for your network
monitoring tool, Jared?  How does it compare with bigbro?  How
configurable is it? How easy is it to add new services or change it "on
the fly"?  Does it know about dependencies?

Numerous isps are using it, along with ITD to replace bigbro.
Grab it at ftp://puck.nether.net/pub/jared/sysmon*gz

        You can also send a message with subject subscribe to
syswatch-request@puck.nether.net to get on the list that I make
announcements on.  Tell me if you find stuff you don't
like.  Here's the answers to your questions.

It's faster than bigbro, and i'm working on doing multiple checks at once
in it.
It knows about dependencies.
You can send it a sighup to reload the config file.
It knows how to talk to e-mail to pager gateways.

Funny you should mention ITD.  I don't know that this really belongs
here in this cf (garage perhaps?) - but anyways-- the way we use bigbro
is we have one main bigbro process with the traditional Big Loop (just
like in sysmon).  Most of our services are RX based, however, so what we
actually have is separate threaded daemons that do RX, and update status
files, and these daemons run separately from bigbro.  Bigbro then runs a
small check program that polls the status file, and does some sanity
checking on the status file (ie, was it updated recently).  We don't use
email anywhere in bigbro (too many things can break) but have a directly
attached modem and a small stupid program that dials out.

So - the bad things about bigbro are - it's proprietary code (CAEN
doesn't want to share it), the config file syntax is ugly, and it does
the One Big Loop.  The good thing is, we can invoke small stupid
programs to do all the real work, including status checking, &
notification.  Also, it's easy to make changes in the config file - it
just notices the file has changed, and reads it in.  Separate threaded
daemons to do most of the real work seems to work really well - that
way, timeouts and other weird slowness in one check don't affect other
checks, & also, when debugging a check daemon, if it blows up, it
doesn't break anything else.

This is all integrated into one ugly set of code.  I'm in the process of
doing some major internal changes to the code and adding more checks.
ITD has a machine dedicated to running my software (sysmon.rs.itd.umich.edu)
and i've done a lot of custom developement for them, including x500 and
other checking that they've requested.  Once I get the multiple checks at
once implemented, there will be no free monitoring software that is better
than my program (IMHO).

The reason for that is everything else uses external programs which take
time to start up, and here it's all in one package, it can send 500
icmp echo-request packets to 500 different hosts and keep track of them,
all at once, not needing to spawn ping 500 times.

This response has been erased.

correct.. my tool does more than that, but I wouldn't want everyone remote
pinging grex all the time and monitoring all the other
services on it to see if it's up.  That'd fill up your connection with
icmp-echo replys

This response has been erased.

Response not possible - You must register and login before posting.

- Backtalk version 1.3.30 - Copyright 1996-2006, Jan Wolter and Steve Weiss