Cannot manually start mysql_init_bundle

In a production ussuri cluster (deployed a year ago), all redeployments stopped working around January for no apparent reason. The culprit appears to be mysql_init_bundle which just hangs.
After using the playbook bellow to debug

--- - hosts: controller-0 become: true tasks: - name: Manage step_2 containers using tripleo-ansible by chko block: - name: "Try mysql_init_bundle" include_role: name: tripleo_container_manage vars: tripleo_container_manage_concurrency: 5 tripleo_container_manage_systemd_order: true tripleo_container_manage_cli: podman tripleo_container_manage_config_patterns: 'mysql_init_bundle.json' tripleo_container_manage_config: "/var/lib/tripleo-config/container-startup-config/step_2" tripleo_container_manage_config_id: "tripleo_step2" tripleo_container_manage_debug: true tripleo_container_manage_check_puppet_config: true tripleo_container_manage_valid_exit_code: [0]

I get the following output from puppet which shows (to me only errors that are bening)

https://pastebin.com/7UUYuJ0E

rpms
podman-1.6.4-4.module_el8.1.0+298+41f9343a.x86_64 podman-manpages-1.6.4-4.module_el8.1.0+298+41f9343a.noarch
containers-common-0.1.40-8.module_el8.1.0+298+41f9343a.x86_64 container-selinux-2.124.0-1.module_el8.1.0+298+41f9343a.noarch containernetworking-plugins-0.8.3-4.module_el8.1.0+298+41f9343a.x86_64
container-tools:rhel8 is used although no binaries have ever been updated according to yum history

CentOS Linux release 8.1.1911 Base OS
External Ceph Backend with an OVS vxlan deployment
Cluster is operational.