High-availability solution for web applications and providing of web services.

High-availability is implemented on multiple levels, with server virtualization technologies and HA clustering using multiple physical and virtual servers. Storage is implemented as redundant, HA-aware storage subsystem - network RAID1. Database high-availability is provided using database clustering where 3 database nodes guarantee maximum uptime.

Web applications, web services as well as database access requests are routed over high availability directors to multiple servers. Physical servers are to be located in different data centers and linked with VPN connections utilizing active OSPF-based routing. Using this approach, all kinds of network servers can make use of high availability concept - DNS, mail, web, ftp, web application servers etc.

All physical and virtual servers and web and networks services are actively monitored. Monitored data is available through web user interface.

Tools used: Enterprise Linux, Pacemaker, DRBD, HAproxy, Zabbix, Percona-XtraDB-Cluster, Tinc, Quagga, BIND, Apache, JBoss AS, Exim, Courier-IMAP