GOAL: have a supervisor alternative on Windows systems.
It allows you to:
QOS restart, stop zeoclients with a load balancing (for the moment only pound is implemented).
- during a reboot of a zeoclient, it is first out of the pool distribution, then restart, and check then if it’s ok it back into the pool.
- Users continue to browse the site . You can restarting all zeoclients this cluster. The site is always available during the reboot modulo the restart of the load balancer (two pool of zeoclients are configured)
To purge reverse proxy cache (for now only squid is implemented)
To monitor the cluster (test url, reporting access times, sending mail if critical event, monitoring memory)
To analyze the logs (as regular expressions) to search across the logs of all machines at once.
To repair a cluster when a critical event is detected (monitoring memory error on 500 test url, log file) -> + restart email alert component.
To rotate log of all the log files present on the cluster.
Contents: