I was running release travis scripts for my Java/Kotlin FOSS to publish artifacts to Maven Central via OSS Sonatype Repository. Bu doing so, the default Maven Deploy plugin publishes to a so-called staging repository and then, the artifact can be synced and released to maven central.
Starting from today (at least I observed it today) this procedure stopped working. After some investigations and reading the change-log - here is my theory…
OSS Sonatype Nexus auto creates a staging repository for the build / deploy based on machine name, ip address and group name of the published artifact. They do have a problem for builds running behind firewalls and changing outgoing IP Address between artifact pushes (see https://issues.sonatype.org/browse/OSSRH-5454).
I read, that you changed your infrastructure setup from container-based towards VM based which involved the change of IP Addresses. This fact should not be a problem, but is it possible that you are using a different outgoing connection infrastructure, than it was for the containers?
Do you have any idea, if my theory is correct and how can I avoid that every upload from the Travis infrastructure comes from a different public IP address? Today I got seven staging repositories instead of one created - in fact this means that publishing of artifacts is not possible anymore, because it separates JARs from Source-JARs from JavaDoc files from POM.xmls to different staging repos and closing / releasing is not possible anymore…
P.S. to be more precise I just got the following IP addresses as sources:
here is the build https://travis-ci.org/holunda-io/camunda-bpm-taskpool/builds/466091296