nagios and its plugins

[iinfi1]

hi I have heard a lot about admins using nagios and its plugins for monitoring of servers
is it safe to install it on SLES itanium and x86_64 production systems
is there any specific plugin I sud be using or do all plugins out there do what they claim?
could you please lead me to any specific plugin you might have used for Oracle RAC environment.

thanks

[Julio Yamawaki]

Hi,

It's absolute safe to use Nagios plugins.
There's no specific plugin for Oracle RAC, but, you can use oracle plugin to show database information and you can customize other(s) plugins to show information about grid process.

Regards,

[iinfi1]

thanks for your input.
sles has some nagios packages along with the DVD ... lemme try them
thanks again

[Julio Yamawaki]

Hi,

I use to get Nagios and plugins from nagios.org, because it's up-to-date.
Also, try to install PHP4Nagios. It's an good tool to make some (good) graphics from Nagios data.

Regards,

[Ralph Grothe]

Because I haven't used SLES there isn't much I can say about their repositories.
But in general they should install easily (just using the distro's respective package manager) and run out of the box.
However, if you compile (and possibly package) for your distro from the sources it gives you more flexibility over certain compile and install parameters, and be it only for the install prefix (prepackaged software often isn't relocatable).
With Nagios (as well as with its plug-ins) you must distinguish between the Nagios server and those parts that usually are to be executed on each host monitored by nagios when the Nagios server polls the configured service checks.
For that reason most Linux distros' package repos also make that split in their Nagios packages. E.g. some nagios-common, nagios-server or merely nagios as well as nagios-plugins packages for the server side, and maybe some nagios-nrpe or only nagios-plugins package(s) for the "client" side.
For the server parts the requirements seem obvious. While for what needs to be installed on each monitored Nagios host (if anything at all) pretty much depends on what you wish to monitor for that particular host.
From the official Nagios Plug-ins, which run the actual host and service checks, maybe one half is fit to execute remotely (that is, nothing needed to be installed on the monitored host for service checks which are covered by those plug-ins), and those will be actively scheduled by the nagios server.
About the other half (or maybe a little less) however, needs to be executed locally on the monitored host (e.g. check_disk, check_procs, check_load etc.).
For those it is required to install them on each monitored host, matching of course as binaries to each host's platform (which needn't be SLES or even any Linux at all).
That may sometimes require to compile the Nagios Plug-ins (or at least those of them that you really require) for the respective platform, shouldn't you be able to find some precompiled software package.
Then you need a means of transport to transmit the check results of those local plug-ins from the monitored host to the Nagios server.
This could either be nrpe and check_nrpe, or check_by_ssh for active checks from the Nagios server, or nsca and send_nsca for passive checks (from the Nagios server's perspective).