A failover occurs in two situations:
- The first situation is when the hardware or operating system has suffered a major failure and the server is no longer functioning. The LifeKeeper core on the backup server detects this when its heartbeat messages fail. At that time, the LifeKeeper core invokes the kit’s recovery script. The recovery script ensures that the Internet server(s) is brought In Service on the backup server.
- The second situation is when the Recovery Kit’s Deepcheck and Quickcheck scripts detect failures of the application. These scripts respond to the LifeKeeper core with code(s) indicating failure. The LifeKeeper core starts the failover process and invokes the kit’s recovery scripts. The LifeKeeper core first stops the local server (if it is not stopped) and deactivates the switchable IP address and LifeKeeper volumes. It then continues the failover to the backup server.
When the primary server is repaired, the Internet server automatically returns to the primary server if switchback type is Automatic. You must perform a manual switchover if switchback type is Intelligent.
A relatively small number of web clients will experience a problem whenever a switchover or failover occurs. First, the process of moving the IP addresses and volumes to the backup system and starting the backup server there takes approximately 45 seconds (depending on the number of IP addresses and volumes), and users cannot connect to the server during that time. Second, the active server is stopped during the switchover, and users with open connections to the active server will be disconnected. In any case, if the IIS client retries the request, the request should succeed. Of course, once the switchable IP addresses and volumes have moved to the backup system and the backup site is running, service will be normal again.
Post your comment on this topic.