[afnog] helpdesk and network management system
Phil Regnauld
regnauld at x0.dk
Fri Aug 19 19:28:57 EAT 2005
On Fri, Aug 19, 2005 at 12:05:12PM -0400, Hannigan, Martin wrote:
>
> Hi, what kind of network managment? Up/down and generic ticketing
> or statistics, or actualy snmp puts? I'd go slow working towards
> equipment control with nms so here's some generic advice.
Nagios. Zabbix seems promising, but since it hasn't earned
its wings yet go for Nagios.
As Randy pointed out, it does scale well (to a certain limit --
we pushed it to 2600 nodes and 5000 service checks, and it choked
when lots of services went down. Though I expect most people
won't be up against this many hosts normally, and you can work around
the problem by spreading the load/sites on multiple boxes).
Nagios is lighweight, but quite advanced, very modular, and extremely
easy to integrate with other tools.
> MRTG: An snmp based performance management system i.e. collects
> statistics via SNMP and keeps history. Highly customizable and will
> take any OID with 2 input values.
>
> RRD Tool: A database to add in multiple flow correlation and other
> highly customizable functions to MRTG.
Take a look at CACTI, for that matter:
http://www.cacti.net/
> Ticketing: Im not up to speed on the latest free ticketing systems,
> but I've used "req" in the past and I know of some large networks that
> still do use it so you may want to look there.
RT. It's easy to install using FreeBSD (using the ports),
it's Perl, and it's easy to customize. Very powerful,
but get a good fast box to run it on. There are a number of add-ons
like RTFM (knowledge-base module) which makes it even more
interesting.
It's possible using Nagios service handlers, to automatically create
tickets / update tickets in RT via the provided API.
> If you are looking for the simplest up/down manager in the world, I
> personally like a tool called "NOCOL", the ncurses version. This is
> supported by linux. It's simple, straightforward and basically says
> "Up" "Down" with no frills, but highly reliable. If I told you the
> name of two large telecom companies that started their networks using
> this tool, you'd be surprised. :)
No surprise, but if they started today, they'd most likely go
for Nagios :)
P.
More information about the afnog
mailing list