Vertx-Hazelcast LocalAddressRegistry messages

[Jan Willem van Markus]

Hi,

We've recently upgraded our application from vert.x 3 to vert.x 4:

Previous configuration:

hazelcast 3.12.13
hazelcast-aws 2.4
vertx 3.9.15

New configuration:

hazelcast 5.3.0 (hazelcast-aws is now part of hazelcast)
vertx 4.5.7

The application is operational both locally and in our AWS cloud environment.

However, when deploying new (EC2) instances, we see a lot of warnings in a short timeframe resulting from Hazelcast, for example:

Sep 12, 2024 @ 10:11:23.423 [hz.optimistic_liskov.IO.thread-in-0] 2024-09-12T10:11:23.322+0200 WARN [com.hazelcast.internal.server.tcp.LocalAddressRegistry]  [10.2.128.124]:5701 [sdbplanning] [5.3.0] Address: [127.0.0.1]:5701 is previously registered with the member uuid: a8f9d858-e9bc-43b8-87d6-e36b2d6e508b to our addressToMemberUuid map, now registered with/overridden by a new member uuid: 0bdd03d0-3c52-4c78-ad30-a3e5eb7091fd.

We expect our cluster is unstable during deployments / autoscaling as a result of the overrides in the LocalAddressRegistry.
We have configured the minimum required parameters as mentioned on the Vert.x website. (default-cluster.xml)

Are above warnings harmless, or indeed an indication of incomplete configuration? We wouldn't expect the member-uuid's to change when an additional instance is added. Also we find it peculiar that the localhost loopback 127.0.0.1 is mentioned as Address in the warning.

Any help is apprciated

(Duplicate of this message is available on Discords (https://discord.com/channels/751380286071242794/751398153420406795)