Configure

Key: application.cluster.autoreset.on.split.brain

Description:
Defines if a TPO should reset itself or not when a split brain situation is detected.

If set to true, both nodes (or the node with the highest order, see "application.cluster.autoreset.on.split.brain.kill.both" description) that run the same group id because of a split brain will shutdown and reboot. After the reboot, the cluster will be in a clean situation again.

When this key is activated, the effect of "application.cluster.lan.checkunplugged" is disabled to avoid conflict ("application.cluster.lan.checkunplugged" put the node in IDLE state when it loses all connections).

/!\ Warning : calls hosted on nodes in split brain situation might be lost as soon as the split brain is detected.

Constraint:
false|true

Related settings:
Auto-reset on split brain buffer time (application.autoreset.on.split.brain.buffer.time)
Auto-reset on split brain - Kill Both (application.cluster.autoreset.on.split.brain.kill.both)
Auto-reset on split brain - Kill Highest (application.cluster.autoreset.on.split.brain.kill.highest)
Check for split brain before DNS table update (application.dns.splitbraincheck)

Key: application.autoreset.on.split.brain.buffer.time

Description:
Defines the time between the detection of a split brain and the reset of the node when "application.cluster.autoreset.on.split.brain" is activated.

By default, the value is 5 seconds.

Related settings:
Auto-reset on split brain (application.cluster.autoreset.on.split.brain)
Auto-reset on split brain - Kill Both (application.cluster.autoreset.on.split.brain.kill.both)