[train] After successful minor update galera will fail to restart

[Michele Baldessari]

After a successful 16.1->16.2 update the overcloud has every entry twice in /etc/hosts. So we have:
# START_HOST_ENTRIES_FOR_STACK: overcloud
172.17.1.60 compute-0.redhat.local compute-0
172.17.3.34 compute-0.storage.redhat.local compute-0.storage
172.17.1.60 compute-0.internalapi.redhat.local compute-0.internalapi
.....
10.0.0.150 overcloud.redhat.local
# END_HOST_ENTRIES_FOR_STACK: overcloud
# BEGIN ANSIBLE MANAGED BLOCK
172.17.1.60 compute-0.redhat.local compute-0
172.17.3.34 compute-0.storage.redhat.local compute-0.storage
172.17.1.60 compute-0.internalapi.redhat.local compute-0.internalapi
172.17.2.135 compute-0.tenant.redhat.local compute-0.tenant
192.168.24.20 compute-0.ctlplane.redhat.local compute-0.ctlplane

So we have all entries under END_HOST_ENTRIES_FOR_STACK:overcloud and then again all entries under "BEGIN ANSIBLE MANAGED BLOCK"

galera calls a getent hosts and a two-line output trips it up and it will error out with:
2021/06/11 15:35:43 socat[1752] E exactly 2 addresses required (there are
3); use option "-h" for help ?

The change that changed the markers in /etc/hosts was https://review.opendev.org/c/openstack/tripleo-ansible/+/788100

We need to remove the old markers.

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: stable/train
Review: https://review.opendev.org/c/openstack/tripleo-ansible/+/796044

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.opendev.org/c/openstack/tripleo-ansible/+/796044
Committed: https://opendev.org/openstack/tripleo-ansible/commit/e3fefe284084aa5e8c14bcf274d0c4d560a476a3
Submitter: "Zuul (22348)"
Branch: stable/train

commit e3fefe284084aa5e8c14bcf274d0c4d560a476a3
Author: Michele Baldessari <email address hidden>
Date: Fri Jun 11 18:20:59 2021 +0200

    [Train-only] Remove duplicate /etc/hosts entries

    After a successful 16.1->16.2 update the overcloud has every entry twice in /etc/hosts. So we have:
     # START_HOST_ENTRIES_FOR_STACK: overcloud
     172.17.1.60 compute-0.foo.local compute-0
     172.17.3.34 compute-0.storage.foo.local compute-0.storage
     172.17.1.60 compute-0.internalapi.foo.local compute-0.internalapi
     .....
     10.0.0.150 overcloud.foo.local
     # END_HOST_ENTRIES_FOR_STACK: overcloud
     # BEGIN ANSIBLE MANAGED BLOCK
     172.17.1.60 compute-0.foo.local compute-0
     172.17.3.34 compute-0.storage.foo.local compute-0.storage
     172.17.1.60 compute-0.internalapi.foo.local compute-0.internalapi
     172.17.2.135 compute-0.tenant.foo.local compute-0.tenant
     192.168.24.20 compute-0.ctlplane.foo.local compute-0.ctlplane

    So we have all entries under END_HOST_ENTRIES_FOR_STACK:overcloud and
    then again all entries under "BEGIN ANSIBLE MANAGED BLOCK"

    The reason for this is that we changed marker via
    https://review.opendev.org/c/openstack/tripleo-ansible/+/788100

    Tested in a Train minor update (16.1->16.2) and the hosts is now
    correct (without the patch it would be broken)

    Change-Id: I85525e29cce9f72382486f8a157bdfc7f3d84f77
    Closes-Bug: #1931729

[OpenStack Infra (hudson-openstack)]

This issue was fixed in the openstack/tripleo-ansible train-eol release.