Email Notification if Server is Unreachable for Whatever Reason?

Z

zacdl

[H]ard|Gawd
Joined
Feb 12, 2007
Messages
2,012
Not really a "problem", but I got to thinking...
Sure- I can get email notifications on my phone if a UPS has lost power, or if the server is under heavy loads and whatnot- but how do I tell when the server is unreachable- for ANY reason?
Say the UPS fails, the NIC fails, or any other reason. How can I be notified of it? The server is down, so obviously the email wouldn't originate from it.

That leaves me to wonder- how do you tell? I'm thinking there has got to be some sort of software somewhere that you install on another PC (maybe the PC and the Server) that talks with the server to make sure it is up. If it is unreachable- it issues an email notification.

That of course brings up a problem to my mind- what if that other PC (workstation or whatnot) is at fault? In other words- the server is fine but the client PC's NIC failed? I would get a false alert. In which case I would need a third PC in the mix to verify results of the other two...

Just brainstorming here- anyone have any ideas?
 
Nagios.
It can be configured to run several different types of checks against a system. It can do simple pings, check service availability (ssh, ftp, http, etc..), or even check if a particular system process is running.

In addition to telling it what to check, you can setup what you would like it to do if a check fails. An easy option is to have it e-mail an individual or group. If you want to be in the know about failures immediately, you could have it mail it to a pager or phone.

It has some gnarly conf files, but most of that is because of how modular/extensible it is designed to be. Things are intimidating at first but that wears off quickly after you get a feel for what conf file does what.
Good stuff.

edit: to answer the deployment question - I guess it's about how anal you want to be about it. I find that having 1 nagios box is fine. If a system has issues then I would be notified of it, if the nagios box has issues then it it would sit until I check on it in the morning. Worst case would be if the nagios box and the monitored system had an issue at the same time.
If you want to get into anal mode, then you could have multiple nagios boxes at different parts of the network/intarweb to check on your monitored server and the other nagios box(es) for more coverage. I personally think that's overkill for the systems I maintain, but could see it if you cared that much about your stuff.
 
Can't find anything on hpov.

Nagios- does that run on Windows? And where do you install it at? Any other workstation on the LAN (like the PC on my desk?)?
 
Nagios is a *nix app. Which I guess could be a deal breaker if you don't know linux nor have any desire to do so.

I would put it on a separate system on the LAN that doesnt get many interruptions (since its job is to sit there and poll things all day.) It really doesn't take much of a system to run it since most of the checks are pretty basic.
 
Nagios is worth learning linux over.

We use it at work. We also have "escalations" where if a problem is not solved in X minutes it actually pages us (for example: incase e-mail/internet is down).

If I had to guess, hpov is short for "HP OpenView"
 
I am planning on eventually adding perhaps FreeNAS on our old server as a backup server... Tons of documents are going to be added to the main server, which won't fit on tape.
Is it able to run on FreeNAS?
 
If your servers are public facing then you can always use something like Hyperspin.com or Pingdom.com, works a treat with email and SMS notifications.
 
I have a simple way, its a but crude but it does me. I connect them all to MSN, if one goes down it no longer shows on MSN.
 
I've had decent luck here with the free version of Big-Brother. http://www.bb4.org/home1.html

It's not the most elegant solution by far, but it was quick to setup and pretty stable. After seing "The Dude" I'm looking at making the transition.
 
zacdl said:
Can't find anything on hpov.

Nagios- does that run on Windows? And where do you install it at? Any other workstation on the LAN (like the PC on my desk?)?
Click to expand...

HPOV = HP Open View, but its pretty costly for licenses. Also look at cacti its free and itsn't rebbly difficult to set up.

A chat program isn't the smartest way to do it. With all the IM epxloits and worms that go arround.
 
I've seen Nagios up and running and must say it rocks. Not to scare you away, but I've been trying to figure out the damn thing for about 4-5 weeks. I've spent most of that time just reading about Linux.

It'll be worth it, once I get it up and running.
 
Another vote for Nagios. Worked for a locla fiber ISP and man Nagios was awesome. We actually had a monitoring server on site, and ANOTHER server off site in another location with beefy battery backup. That way if the on site server went down, generator didnt kick on, fiber was cut/burned, etc the second site would let us know.

It is pretty gnarly to setup, but worth it, as stated above. My vote is in Nagios.
 
I can't believe I've never heard of Nagios...and here I paid several thousand dollars for Solarwinds and Lansurveyor.... (not that either of those suites are bad, because they completely rock...but man...time to put up another linux box).

OT: If you have the money and don't want to tool around in linux, Solarwinds is one of the most complete network monitoring solutions I've ever seen. Check it out.
 
You can configure alerting in Microsoft System Center Operations Manager. You can also generate SMS alerts as well.
 
You must log in or register to reply here.
Back
Top