When you configure application monitoring, ApplicationHA uses a heartbeat to communicate the application status to VCS in the virtualization infrastructure layer. If the application or its component fails, ApplicationHA attempts to restart it. If the application does not start, depending on the configuration, ApplicationHA initiates a soft reboot of the managed LPAR. If the application does not come online after the soft reboot, ApplicationHA stops sending the heartbeat to VCS. Depending on the configuration, VCS in the infrastructure layer then performs a hard restart of the managed LPAR.
A hard restart has various implications and may not be the desired solution at all times. It may prove to be counter productive in several cases, including the following:
The managed LPAR itself is running fine but the application is unable to get the required resources
The other applications and tools running on the machine may either hang or take longer time to recover or restart after an abrupt shut down
A hard restart can be potentially disruptive if there are multiple critical applications running on the managed LPAR
A soft reboot or an operating system driven restart is graceful and allows for a more orderly shutdown of applications and tools running on the machine and can help reduce potential disruption to critical applications. ApplicationHA provides this layer of application control wherein you can configure ApplicationHA itself to restart the managed LPAR using native operating system commands. Note that the soft reboot of the managed LPAR by ApplicationHA is disabled by default.
ApplicationHA provides attributes that you can use to configure ApplicationHA to restart the managed LPAR.