Download presentation
Presentation is loading. Please wait.
1
System Monitoring With Nagios in CSCF
Very very general overview - nothing new to anybody who attended Patrick’s WatITIs talk, although it covers some of the general ground I think he missed. Picture credit: (CC licensed) This document is the IP of the University of Waterloo, and was entirely written by on his own time. Everything is worth what you pay for it, and no more, so caveat emptor and all that. Presented to CSCF staff at a group meeting, 24 May 2006.
2
What Is Nagios? Succinctly, a set of server scripts, CGIs, and other programs that monitor systems and services on client hosts. Clients can be anything with a network connection and a TCP stack. The CGIs provide a web-based interface. Backend is a C program
3
What did we do before? “Just In Time”, when we were lucky.
4
Firefighting We usually found out something was wrong from somebody else: students, faculty members, or we’d get bouncing. Sometimes that somebody else would be angry that we hadn’t noticed before, leading to a different kind of flame. (picture source)
5
What about now? Simple monitoring for things like basic network connectivity (ICMP echo)…
6
Monitoring the bunnycam
Monitoring the bunnycam! Checking for slightly more complex things like server response time.
7
Simple temperature and humidity sensors
Simple temperature and humidity sensors. Came in handy last night (23rd) as PLG’s machine room started overheating. Grad student noticed it too, but… this gave us some proof that we could give to Plant Operations.
8
Where do we want to be?
9
Ideally, we’ll know something’s wrong - and maybe even fix it - before the users notice.
10
What can it do? It can glue things together. Plugins can be written in any language, and with clientside permissions we can run anything there too. Like hot glue though, there’s limits: it’s not great at SNMP - doesn’t catch traps. Also not the easiest thing in the world to configure non-trivially. Maybe we can’t get HAL’s near omnipotence, but we can get close. It’s possible to store host and service performance data, but we don’t currently do that. Included nagiostats utility will write MRTG output for pretty graphs too. Picture credit: World Precision Instruments,
11
References http://www.nagios.org - project homepage
- plugin exchange site RTs: 51804, 49405, 47625/47698, 51585, Also for MFCF side. - uwdir credentials Nagios.cscf is a Sun running FreeBSD Not xhier’ed, but it is set up to operate basically as I described a few years ago in my WatITIs presentation on operating system upgrade management using FreeBSD. Xhier project was active, but is stalled now.
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.