Monitoring a bunch of PCs for online status and things

[Kelemvor]

Howdy.

I'm looking for a way that I can monitor about 2500 PCs that I am in charge of so that I can tell which ones are online and which are not by looking at a nice chart or something.

I was trying out Spiceworks, which was recommended to me by someone, and it looks great with one problem.

The IP addresses of the machines I need to monitor are not in a nice range. They are spread out over the whole network and all end in 221 or 222.

For example:
1.1.1.221
1.1.1.222
1.1.2.221
1.1.2.222
1.1.3.221
1.1.3.222

And this would continue through 5 full cycles all the way up to 1.5.254.222. There's no easy way to write this as a standard range, so I'm not sure what to do. I basically need a program that can import the set of IPs that I have and use that instead of scanning the network to find the devices.

Any ideas for things I can check out would be great.

Thanks!

 

[drebo]

Step 1: fire whoever came up with that IP numbering scheme.

Step 2: specify what kind of monitoring you can do. (Are they windows machines, can you install SNMP on them, what do you need to monitor (uptime only or do you need specific stats about them), etc, etc)

Step 3: Find whatever program you need to do whatever it is you need to do. For simple ping monitoring, Whatsup Gold might work. If you need more, you might need to go with something SNMP or WMI based, like Spiceworks.

 

[DnetMHZ]

Step 1: pistol-whip whoever came up with that IP numbering scheme.

There, fixed that for ya.

 

[bobdole369]

Basic network monitoring such as: SpiceWorks was made for this sort of thing. Tell it to scan the entire 10 subnet or whatever and it'll eventually get them all.

In spiceworks you could specify that 221 and 222 for every cotton picking network known to man, but that would take literally years. Seriously find that guy and punch his mama in the face.

SW can handle huge IP ranges like that without tooo much trouble. You just set it to scan and off she'll go.

 

[Kelemvor]

I know it sounds screwy but here's the setup. We have over 1000 locations across the country. Each location has it's own network. Each network has two machines I need to monitor. They are all setup with an IP around the 221-223 range, but the first 3 numbers of the octet are different for all 1000 locations. That's why it's such a screwy range I need to scan.

And these are just Win XP machines doing specific tasks. Just looking for basic info such as if they are online or not, maybe when they rebooted last, etc.

I did find a way to get spiceworks to scan what I need, but SW's support people told me they only recommend their program scan 1000 devices per install. So I'd need to install it on 3 machines to scan and have them all feed back into a primary one.

I think I'll just set the whole thing up and see what I end up with.

 

[yinan]

SCCM

 

[mammador]

Second firing the person who devised that IP addressing scheme. lol..

Invest in a decent NMS.

 

[High Lord Gomer]

If you have that list of devices in a text file, edit it to look like a hosts file and then you can have WhatsUp do a host file import to get them all in quickly without scanning the rest of your network.

 

[Kelemvor]

Well yeah but it looks like Whatsup costs thousands of dollars so that's not going to happen here.

 

[ViviTheMage]

I am surprised no one mentioned nagios. It will do this for you .

Why fire the guy who did the IP naming? Sounds like they're on different subnets...granted, servers generally start in the .1-.20 range.

 

[drebo]

If you just need simple uptime monitoring, smokeping will do it for you. You have to manually create the config files, but a simple excel macro should do that for you without too much work.

 

[Brovane]

Do you have any budget? How do currently manage the PC's for updating the OS, software?

 

[Kelemvor]

We have no budget. Other teams manage the PCs for updates and things. We just manage the apps that run on the PCs to run the front end. We're just trying to find a way that we can tell when something isn't working since other teams don't really monitor it that closely on a daily basis.

 

[Brovane]

We have no budget. Other teams manage the PCs for updates and things. We just manage the apps that run on the PCs to run the front end. We're just trying to find a way that we can tell when something isn't working since other teams don't really monitor it that closely on a daily basis.

Yeah that is typical IT management. Monitor this but we have no budget. Remember nothing is free even software that says it free because there is still soft costs to getting the software up and running.

 

[drebo]

We have no budget. Other teams manage the PCs for updates and things. We just manage the apps that run on the PCs to run the front end. We're just trying to find a way that we can tell when something isn't working since other teams don't really monitor it that closely on a daily basis.

You still have not told us what you're trying to monitor. Several options were given which monitor various different things. You should be able to piece something together at this point.

 

[bobdole369]

This screams nagios. All you need is uptime (maybe a port) - it WILL take a bit on the frontend to set it up (your time mostly designing scripts), but will alert you when one is down, and will display nice graphs (again with more work).

You will need to dedicate at least a VM to it (couple gigs space, 256-512MB RAM), Linux based distro (CentOS my personal fave for it).

 

[Brovane]

This screams nagios. All you need is uptime (maybe a port) - it WILL take a bit on the frontend to set it up (your time mostly designing scripts), but will alert you when one is down, and will display nice graphs (again with more work).

You will need to dedicate at least a VM to it (couple gigs space, 256-512MB RAM), Linux based distro (CentOS my personal fave for it).

The problem from what I can see is that it looks like he also needs to see uptime for applications on the PC's. This tells me there is going to be sometype of WMI or SNMP pulling on the individual PC's. I can do that real easy using Orion APM but this solution is way outside the budget of $0.