**Please read - Simple Network / PING Monitoring**

[VFaretrajr]

I hope this post serves as a dual-role benefit to those reading it. Especially as a relative newbie to Cacti & Linux programming.

The first being to benefit on how to utilize Cacti perform a - what I perceive to be - simple task of network monitoring.

The second is so that those following can learn from the (Hopefully) numerous posts that will go on with this thread.

OK - Here goes nothing!

The Scenario:

You have a WAN network consisting of one workstation, at one location and you have 1200 locations.
The IP addresses at each location is static and the workstations are nothing more than just operating as a standard network share for local users - Nothing more, nothing less. The only requirement for these workstations is that they are NOT to be powered off / taken off the network at ANY given point in time. They are to remain online 24x7x365.

The problem:

At the present time, there is NO monitoring what-so-ever on these devices as to when they go down. Furthermore, there is NO logging of when the device went down (and if / when it came back up) to catch the "repeat offenders". Some scenarios include circuit breakers being tripped, office relocation, cleaning within the office disconnected network cable, uplink switch failed, etc.

The solution:

I would like to utilize a network monitoring software that does the following:

1. I would like to utilize a “PING” like function as the only method to see if the host is alive.
2. Polling can be done every 10 minutes as this is not a true mission critical device.
3. There should be at least 2 or 3 attempts to await for a reply from the device before reporting a "DOWN" state.
4. If the site times out, an email alert should be emailed as well as for when the site comes back online.
5. Provide a "real-time" web view of ALL devices whether offline or online.
6. Lastly, there should be a log file which retains all of the up / down status changes for ALL these sites to locate the repeat offenders.

In short, a simple network monitor without utilizing all the bells and whistles that SNMP offers. Let's be honest, at the end of the day - we as network administrators can be as proactive as possible to avoid downtime, however, that creates for this perception that nothing should EVER go down.

In my opinion, while that is true - we must be reactive and responsive as to when it does go down and addressing the issue immediately.

I hope this post creates quite a stir & following to learn from and get some positive feedback.

Thanks,
Vince

[micke2k]

Use Nagios for this. It fills all the functions you need. On the repeat offenders/links you will use cacti + advanced ping.

Nagios will check up/down for you and email. Cacti will help you troubleshoot bandwidth/latency issues.

Download CactiEZ and you'll have both up and running on the same box in 10 min for testing. For production i would deploy two different servers.

Then you can script in all your hosts. You probably have all the ips in an excel sheet. Should already be scripts out there for this.

Edit:

I have done this on 1600 hosts on nagios. These kinds of environments quickly deteriorate and produce a humongous amount of alerts.

Image the internet/mpls links going up down multiple times a day. I would limit it to check during certain hours, between 9-5 as an example. Aswell as increasing the time for it to send an alert.

Also monitor the routers, not the hosts. Hosts will go up/down all the time.