Can a Zookeeper cluster deployed with three pods recover itself when two pods are deleted? If it has this capability, how is it achieved?

[NJUSTER-lxw]

Description

Can a Zookeeper cluster deployed with three pods recover itself when two pods are deleted? If it has this capability, how is it achieved?

What I'm observing now is：
When only one pod exists, it is configured for three nodes, which causes the cluster to fail to reach a quorum of two or more nodes. Therefore, when the pod is restored, the reconfiguration operation in the startup script cannot be responded to, causing the cluster to fail to recover.

Importance

Need the ability to recover from abnormal situations

Location

Suggestions for an improvement