instack-undercloud manifest can conflict with systemd puppet module

[Michele Baldessari]

While rebasing puppet-rabbitmq due to other reasons, I noticed that the undercloud (re) deploy would fail with:
2019-05-31 09:59:00,285 INFO: ^[[1;33mWarning: ModuleLoader: module 'zaqar' has unresolved dependencies - it will only see those that are resolved. Use 'puppet module list --tree' to see information about modules
2019-05-31 09:59:00,285 INFO: (file & line not available)^[[0m
2019-05-31 09:59:00,919 INFO: ^[[1;31mError: Evaluation Error: Error while evaluating a Resource Statement, Evaluation Error: Error while evaluating a Resource Statement, Duplicate declaration: Exec[systemctl-daemon-reload] is already declared in file /etc/puppet/manifests/puppet-stack-config.pp:99; cannot redeclare at /etc/puppet/modules/systemd/manifests/systemctl/daemon_reload.pp:5 at /etc/puppet/modules/systemd/manifests/systemctl/daemon_reload.pp:5:3 at /etc/puppet/modules/rabbitmq/manifests/config.pp:231 on node undercloud-0.redhat.local^[[0m
2019-05-31 09:59:00,982 INFO: + rc=1

This is because a newer puppet-rabbitmq introduced:
+
+ if $::facts['systemd'] {
+ Class['systemd::systemctl::daemon_reload'] -> Service['rabbitmq-server']
+ }
   }

Which then pulls in the exec called 'systemctl-daemon-reload' from the systemd module which conflicts with the same call we do in undercloud puppet manifest

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: stable/queens
Review: https://review.opendev.org/662488

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.opendev.org/662488
Committed: https://git.openstack.org/cgit/openstack/instack-undercloud/commit/?id=d12465b4a0134a2ecc0d5e5d9fb96cf55a730b6e
Submitter: Zuul
Branch: stable/queens

commit d12465b4a0134a2ecc0d5e5d9fb96cf55a730b6e
Author: Michele Baldessari <email address hidden>
Date: Fri May 31 16:33:34 2019 +0200

    [queens-only] Rename the systemctl-daemon-reload exec

    While rebasing puppet-rabbitmq due to other reasons, I noticed that the
    undercloud (re) deploy would fail with:

    2019-05-31 09:59:00,919 INFO: Error: Evaluation Error: Error while evaluating a Resource Statement, Evaluation Error: Error while evaluating a Resource Statement, Duplicate declaration: Exec[systemctl-daemon-reload] is already declared in file /etc/puppet/manifests/puppet-stack-config.pp:99; cannot redeclare at /etc/puppet/modules/systemd/manifests/systemctl/daemon_reload.pp:5 at /etc/puppet/modules/systemd/manifests/systemctl/daemon_reload.pp:5:3 at /etc/puppet/modules/rabbitmq/manifests/config.pp:231 on node undercloud-0.redhat.local
    2019-05-31 09:59:00,982 INFO: + rc=1

    This is because a newer puppet-rabbitmq introduced:
    +
    + if $::facts['systemd'] {
    + Class['systemd::systemctl::daemon_reload'] -> Service['rabbitmq-server']
    + }
       }

    Which then pulls in the exec called 'systemctl-daemon-reload' from the
    systemd module which conflicts with the same call we do in undercloud
    puppet manifest.
    Let's just rename this exact to 'systemctl-daemon-undercloud' so we do
    not duplicate this resource ever. Tested this on a queens undercloud
    with a puppet-rabbitmq that triggers the problematic exec and got a
    correctly redeployed undercloud.

    We could have used the systemd module instead but that might require
    other backports and this seems the smallest possible approach to solve
    this.

    Change-Id: I96f5e4e51947da6919724d8ffb8531b8ecf87ff8
    Closes-Bug: #1831253

[OpenStack Infra (hudson-openstack)]

This issue was fixed in the openstack/instack-undercloud 8.4.8 release.

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: stable/rocky
Review: https://review.opendev.org/673711

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.opendev.org/673711
Committed: https://git.openstack.org/cgit/openstack/instack-undercloud/commit/?id=764aaf6a5cfc6906e673322fac82d7b58419907e
Submitter: Zuul
Branch: stable/rocky

commit 764aaf6a5cfc6906e673322fac82d7b58419907e
Author: Michele Baldessari <email address hidden>
Date: Fri May 31 16:33:34 2019 +0200

    [queens-only] Rename the systemctl-daemon-reload exec

    While rebasing puppet-rabbitmq due to other reasons, I noticed that the
    undercloud (re) deploy would fail with:

    2019-05-31 09:59:00,919 INFO: Error: Evaluation Error: Error while evaluating a Resource Statement, Evaluation Error: Error while evaluating a Resource Statement, Duplicate declaration: Exec[systemctl-daemon-reload] is already declared in file /etc/puppet/manifests/puppet-stack-config.pp:99; cannot redeclare at /etc/puppet/modules/systemd/manifests/systemctl/daemon_reload.pp:5 at /etc/puppet/modules/systemd/manifests/systemctl/daemon_reload.pp:5:3 at /etc/puppet/modules/rabbitmq/manifests/config.pp:231 on node undercloud-0.redhat.local
    2019-05-31 09:59:00,982 INFO: + rc=1

    This is because a newer puppet-rabbitmq introduced:
    +
    + if $::facts['systemd'] {
    + Class['systemd::systemctl::daemon_reload'] -> Service['rabbitmq-server']
    + }
       }

    Which then pulls in the exec called 'systemctl-daemon-reload' from the
    systemd module which conflicts with the same call we do in undercloud
    puppet manifest.
    Let's just rename this exact to 'systemctl-daemon-undercloud' so we do
    not duplicate this resource ever. Tested this on a queens undercloud
    with a puppet-rabbitmq that triggers the problematic exec and got a
    correctly redeployed undercloud.

    We could have used the systemd module instead but that might require
    other backports and this seems the smallest possible approach to solve
    this.

    Change-Id: I96f5e4e51947da6919724d8ffb8531b8ecf87ff8
    Closes-Bug: #1831253
    (cherry picked from commit d12465b4a0134a2ecc0d5e5d9fb96cf55a730b6e)

[OpenStack Infra (hudson-openstack)]

This issue was fixed in the openstack/instack-undercloud 9.5.1 release.