The main back-end processing relies on three CRON jobs that run a various times:

*/5 * * * * /home/cactiuser/scripts/eumetsat_utilisation.sh > /dev/null 2>&1
*/2 * * * * /home/cactiuser/scripts/dfn_monitoring.sh > /dev/null 2>&1
*/1 * * * * /home/cactiuser/scripts/cluster_check.sh > /dev/null 2>&1 

*** cluster_check.sh only runs on the Backup Server prod-eumetsat02.geant.net

cluster_check.sh

This calls a PERL script /home/cactiuser/scripts/cluster_check.pl using a single parameter of the IPv4 Address of the primary production instance (62.40.120.236).

In a nutshell, the script tries to Ping the primary server 3 times. If it does not get a response, then it checks for the existence of the  /var/lib/dbeacon/EUMET/failover.chk file. If it does not exist, then it creates the file with the contents Process Emails And Alarms. The existence of this file tells the Backup server that it needs to take control of the back-end monitoring.

dfn_monitoring.sh

eumetsat_utilisation.sh

MySQL Database Connections

The Eumetsat back-end processing connects to a MySQL database:

dbeacon_cacti database on eumetsat.geant.net using user eumetsat