Strange state "UNKNOWN"

[Schindewolf, Stefan, Infraserv-Hoechst/DE]

Hello Nagios Users.

I am experiencing a strange problem with the "check_by_ssh" plugin. Nearly
3/4 of my services are check by this plugin.

However, some services go into "unknown" state and flap around between
"unknown" and "ok".
The message output of every "unknown" service indicate an "ok" state.
When I hit "Reschedule" and "Force Check", the services allways go into
"OK", but later return into "unknown".
I can not recognize a regularity, the problem seems to occur accidently.

I supposed a "return code" problem. So I tried to run the check in the shell
as user "nagios". Everything is ok, message is displayed correctly, even the
return code ("echo $?" gives the return code) indicates "OK" ("0" in the
Shell).

Unfortunately Nagios does not seem to catch it. I wrote a perl script, which
executes the checks for nagios, so that nagios virtually runs a local
command. This script executes "check_by_ssh" plus a check command. Same
Result. The return code is 0. The message output is like "OK! ..."
Nagios says "unknown".
Using the "real" ssh command for check execution does not change anything.
I tried to use nrpe, but to compile and configure it on up to 40 hosts seems
very time consuming.

At the moment I have no further idea, what to do and would like you to
provide some advice.

Thanks and best regards.


Stefan Schindewolf

Infraserv GmbH & Co Höchst KG
Service Center Informationstechnologie
D710, D-65926 Frankfurt
Telefon: (069)305 - 83010
Fax: (069)305 - 23549
Mail: stefan.schindewolf at infraserv.com


-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf