An active/passive configuration is the most basic HA configuration. In this type of configuration, one node[1] is active running a ResourceGroup[2], and the other node is passive, simply waiting for the active side to fail.
The advantages of this configuration are that failing over causes no service degradation, provided that the passive machine is similar to the active machine. In addition, services are only restarted when machines fail, which typically results in half as many services interruptions as an active/active configuration.
The disadvantage of this configuration is that the passive machine isn't providing any kind of useful service, and is sometimes viewed as an inefficient use of resources.
To instruct Heartbeat[3] to operate in active/passive mode, simply make sure you configure the auto_failback[4] directive in your ha.cf[5] file like this:
auto_failback off
If you wish to fail resource[6]s back to failed machines manually, then configure your system as an active/passive arrangement, and use the hb_standby[7] command to move resources back to their preferred machines[8] under administrator control.
| [1] | http://www.linux-ha.org/node |
| [2] | http://www.linux-ha.org/ResourceGroup |
| [3] | http://www.linux-ha.org/Heartbeat |
| [4] | http://www.linux-ha.org/ha.cf/AutoFailbackDirective |
| [5] | http://www.linux-ha.org/ha.cf |
| [6] | http://www.linux-ha.org/resource |
| [7] | http://www.linux-ha.org/hb_standby |
| [8] | http://www.linux-ha.org/PreferredNode |
| [9] | http://www.linux-ha.org/ActiveActive |
This information provided courtesy of the Linux-HA project at http://linux-ha.org/