On Tue, 2004-05-11 at 16:22, Roland Venter wrote: > Hi Folks, > > I'm trying to put together a shell script to monitor several Linux servers, > Redhat, SUSE and Fedora > > A quick google search reveals loads of utilities, etc. I'm constrained by > the > following: > > 1. I do not 'own' some of the remote servers and only perform monthly > server checks and compile a report with recommendations, etc > So I can't install, make any changes to exiting configuration, compile code > or modify core files without proper change control. > > 2. Some of the remote servers are not visible from the outside, only > incoming SSH, this rules out any client/server configuration or SNMP > > I've been running logwatch each night to send a consolidated log report. > > Presently I'm using cron to run a shell script do give a system overview of: > > last | grep reboot : check for reboots > cat /proc/mdstat : RAID Status > uptime > /proc/meminfo : Memory Usage > netstat -an | grep -c ESTABLISHED : TCP connections > /bin/netstat -tupan > df -h : Disk usage > /usr/sbin/smartctl -a /dev/hda : SMART Status of disks > /usr/sbin/ntpq -p : Time Sync > I'm sure there must some "Health Check" scripts either Bash or Perl on the > net, > which will save me reinventing the wheel. > > Any suggestions or links would be appreciated or if you've done something > similar please let me know > > Thanx, > > Roland I would suggest you use Nagios [1]. In conjunction with some of the available plugins, it'll do what you need and more. All you need is one box to consolidate all the monitoring data, and you can even squeeze some nice and pretty graphs for management :) [1] http://nagios.sourceforge.net Have fun! Cheers, Callan
Attachment:
signature.asc
Description: This is a digitally signed message part
