I need to monitor a compute farm system.  The system is made up of 18 hosts. 
I have inherited a functioning command (example below) 
that gives me what I want from the command line: an overview of the health of 
18 devices in the service.

I would like to know the recommended way (Best Practices) to get two things:
1) I would like this manner of overview of all 18 devices in the Zenoss GUI, and
2) I would like an Error event raised when status on one or more devices goes 
from "ok" to "unavail"

I have reviewed zencommand and ssh modelling in both the Admin guide and the 
forum.
I can model individual devices via SNMP just fine.
The best hint I have seen so far is from this post
http://forums.zenoss.com/viewtopic.php?t=7470&highlight=ssh+command

I thought about doing this by creating a /device/lsf class for this compute 
farm,
wrapping the command below in a shell script check_lsf ,
dropping that shell in /opt/zenoss/libexec/
running the shell script through zencommand, 
then parsing the output in the shell script
and using the output in a new Data Source.
That would raise alerts on "ok" or "unavail"

However, I have a need to display the rows and columns of the command as well, 
either in this form direct from the shell script 
or some other manner more native to Zenoss.
I don't want to call the command twice.

I run the command and I get an overview of the 18 devices.  

[EMAIL PROTECTED] ~$ ssh 134.87.177.20 /tools/bin/lsload
HOST_NAME       status  r15s   r1m  r15m   ut    pg  ls    it   tmp   swp   mem
chew                ok   0.0   0.0   0.0   0%   0.0   0  4728 9248M   32G 7736M
gaff                ok   0.0   0.0   0.0   0%   0.0   0  6024 9288M   32G   15G
kaiser              ok   0.0   0.0   0.0   0%   0.0   1  4604 9296M   32G   15G
sebastian           ok   0.0   0.0   0.0   0%   0.0   0 17760 9296M   32G   15G
tyrell              ok   0.0   0.0   0.0   0%   0.0   0   301 9296M   32G   15G
zhora               ok   0.0   0.0   0.0   0%   0.0   0  5968 9296M   32G   15G
cannonball          ok   0.0   0.0   0.0   0%   0.0   0 15816 9296M   16G 5260M
skyride             ok   0.0   0.0   0.0   0%   0.0   0 79808 9296M   16G 5300M
swanboats           ok   0.0   0.0   0.0   0%   0.0   0 79808 9296M   16G 5292M
roy                 ok   0.0   0.0   0.0   0%   0.0   0 36160 9296M   32G   11G
rachel              ok   0.0   0.0   0.1   1%  20.4   3   237 3634M   31G   10G
batty               ok   0.1   0.0   0.0   0%   0.0   1 54944 9296M   32G   15G
eurobungy           ok   0.2   0.0   0.0   0%   0.0   0 79808 9296M   16G 5516M
holden              ok   0.3   0.1   0.0   0%   0.0   0 45056 9296M   32G   15G
bryant              ok   0.6   1.4   0.8  33%   0.0   1   425 9304M   32G 4582M
pris                ok   0.9   0.0   0.0   0%   0.0   0   126 9296M   31G   14G
taffey              ok   1.0   1.0   1.0   6%   0.0   1    65 9296M   32G   59G
deckard             ok   1.1   1.0   1.0  50%   0.0   1   403 9296M   32G   15G
[EMAIL PROTECTED] ~$

The compute farm system is referred to as Load Sharing Facility, or lsf
from www.platform.com

I expect there is the quick ugly way to do this, and the elegant way.  Given a 
choice, I am today looking for the quick ugly way  :-)  and will bring in the 
elegant way later...

This example is from one site.  There are more sites after this site.

I run Zenoss 2.2.4 on RHEL4

Any help would be appreciated.  
So would a detailed example of how someone else is successfully using 
zencommand.

Thank you in advance,

David Sloboda 
PMC-Sierra, Inc.      
_______________________________________________
zenoss-users mailing list
[email protected]
http://lists.zenoss.org/mailman/listinfo/zenoss-users

Reply via email to