We are encountering a race condition where followers struggle to join the leader within a Raft cluster. Each pod’s cluster address is set to its DNS address. The pod’s liveliness and readiness checks use the “vault status” command, meaning the followers’ DNS names must be resolvable in order for them to join the leader. However, those DNS names remain unresolvable until the liveliness and readiness checks pass, and those checks require the followers to join the leader first. Can someone please help fix this ?
Adding publishNotReadyAddresses: true
in k8s service helped fixed the issue.