[Nagiosplug-devel] [Nagios-users] nrpe issue with solaris 9?

[Hari Sekhon]

Andreas Ericsson wrote:
> About a billion times. You're running the check on the nrpe side as root,
> but nrpe runs it as someone else, and the plugin is crap so it prints
> "ok" when it fails to get the proper numbers. I'm guessing you wrote the
> plugin check_fs_space yourself, and only tested it as root, right?
>   
I think this is a very generic problem among plugins, even those in the 
standard distribution's contrib dir, especially those written in perl,
which is a real shame, cos I try to persuade people at work that people 
who know perl are usually really great, but they insist that python
makes people write better code.

I see plugins here and there of 150 lines of code, with practically no 
error checking, when really it should be at least twice that... they 
look like
they were written in a real rush. I think we should all try to make our 
plugins more rock solid... they should be black box near perfect. I've made
mistakes like everyone else who has written plugins, that's what 
revision numbers are for... ;-), but we should try to do better.


I've just had to upgrade my plugins cos the check_ntp plugin had not 
quite the right output and sent it to stderr so I got No Output in 
Nagios. I've made
this mistake myself before so I know that usually you would do this for 
ordinary programs, but Nagios doesn't catch the stderr.
On this note, perhaps Nagios should catch both stdout and stderr and let 
program write to stderr as is more std in the programming world?

-h