Elastic 8.0 refuses connections using hostname instead of localhost

Hello team,

I would like to ask you for your expertized advice.

After an one of Elastic restart WITHOUT TOUCHING CONFIG :slight_smile: it started refuse connections using

I have tried variations found on this forum even on others.
Nothing working.

This is snipped from my config including commented attempts.

#network.host: czchown5029751.blabla.com
#network.host: [  _site_ , _local_ , _global_ , czchown5029751.blabla.com ]
#network.bind_host: [  _site_ , _local_ , _global_ ]
#network.publish_host: czchown5029751.blabla.com
  bind_host: [  _site_ , _local_ , _global_ ]
  publish_host: czchown5029751.blabla.com

In log I can see:

[2022-11-21T10:50:18,406][INFO ][o.e.x.d.l.DeprecationIndexingComponent] [CZCHOWN5029751] deprecation component started
[2022-11-21T10:50:18,593][INFO ][o.e.t.TransportService   ] [CZCHOWN5029751] publish_address {czchown5029751.blabla.com/}, bound_addresses {}, {[::1]:9300}, {}, {}
[2022-11-21T10:50:19,360][INFO ][o.e.b.BootstrapChecks    ] [CZCHOWN5029751] bound or publishing to a non-loopback address, enforcing bootstrap checks
[2022-11-21T10:50:19,368][INFO ][o.e.c.c.Coordinator      ] [CZCHOWN5029751] cluster UUID [XY3O-wkiTDuP6SKqe47LHA]
[2022-11-21T10:50:19,511][INFO ][o.e.c.s.MasterService    ] [CZCHOWN5029751] elected-as-master ([1] nodes joined)[{CZCHOWN5029751}{NmAi12d5SFS_E8BNDyEqFA}{Y1hIgzKEQ_SNrU_EtYZdnQ}{czchown5029751.blabla.com}{}{cdfhilmrstw} elect leader, _BECOME_MASTER_TASK_, _FINISH_ELECTION_], term: 29, version: 1213, delta: master node changed {previous [], current [{CZCHOWN5029751}{NmAi12d5SFS_E8BNDyEqFA}{Y1hIgzKEQ_SNrU_EtYZdnQ}{czchown5029751.blabla.com}{}{cdfhilmrstw}]}
[2022-11-21T10:50:19,637][INFO ][o.e.c.s.ClusterApplierService] [CZCHOWN5029751] master node changed {previous [], current [{CZCHOWN5029751}{NmAi12d5SFS_E8BNDyEqFA}{Y1hIgzKEQ_SNrU_EtYZdnQ}{czchown5029751.blabla.com}{}{cdfhilmrstw}]}, term: 29, version: 1213, reason: Publication{term=29, version=1213}
[2022-11-21T10:50:20,197][INFO ][o.e.h.AbstractHttpServerTransport] [CZCHOWN5029751] publish_address {}, bound_addresses {}, {[::1]:9200}, {}
[2022-11-21T10:50:20,200][INFO ][o.e.n.Node               ] [CZCHOWN5029751] started

As you can see, it will publish the hostname on port 9300 which I didn't defined but not for port 9200.

Is it a problem between chair and keyboard or really there is a bug?

Thank you for help.

What are you trying to achieve with the differences between bind_host and publish_host?