Shell Programming and Scripting

It feels like I have taken on to of a big task.. but I want to learn but dont know where to start.. Rome wasnt built in one day right..

I have around 30+ servers running linux/unix flavours and need to be able to do this in a script that will be a cronjob.
- ssh to all 30+ servers
- do df -kh
- save server name and result to TodaysDate.log
- check result if diskspace/partition is less then 20%
- send an email if the above is true.


Bare in mind I have never written any advanced scipt.. so this is probably simple for you h4x0rs out there..
In my head
SERVERLIST=servers.txt
SSH first hit from servers.txt
ssh user@server.lan df -kh
save result in Todaysdate.log
move to next server in list
if no more server then check todaysdate.log
if space is less then 20% then
email admin@n00badmin.com
end


not sure I will be able to do ssh keys on every machine.. so an option would be to save the pwds in a file.. its a closed LAN.. its sounds bad I know but if I could start with that..

Why dont you use Nagios for this issue? You can install Nagios one Master server that you will choose, and then install Nagios NRPE in all other 29 servers.

In Nagios, NRPE reports remote hosts status to main Nagios server and you can monitor all the Servers on the screen of Master server. Using "du -sh" or "df - kh" you can write an easy script and then put it's path in all 29 NRPE servers' NRPE config files and finaly you collect all the status of those servers. Use Active checks for this purpose. You can set check status of scripts also in config files.

You can also check results of some scripts ( which is placed in crontab) periodically by writing nagios plugin.

I know at first sight it might look so different and diffucult to you but you can do it searching..
I hope i give you a way to do it
Regards

you are correct.. we run Nagios on other systems so Im familiar with that.. but these systems are a big no no.. I cant go into what they do but we arent allowed to install Nagios on them

youre thinking is in the right way though.. Nagios was my first thought too

Yes really, once i needed to monitor 6 production servers and i did know nothing about nagios but searching and searching i learnt enough to configure considering requirements. I offered it because of monitoring is good, and once you unserstand configuration method its so usefull. Im afraid you should use ssh


and in df -kh you can get the % column info like

Code:

df -kh | nawk '{print $(NF-1)}'

or this one give only % usage of root partition "/"

Code:

df -kh | nawk '/^\/ /{print $(NF-1)}'

and use such example values in comparision loop or loops according to critical level in requirement and go on