In case Kolla's users want to deploy with both of
binary and source image, we should have a variable
install type that define install type for each project.

We also add specific image tag for each Openstack project.

This commit is implemented for Designate, Gnocchi and Heat projects.

Change-Id: I07451750e70e0e6305dca451422e33cd31ce8a4c
Implements: blueprint mixing-binary-and-source-image

Fix the following error:
openstack_swift_auth is undefined
unsupported parameter for module: project
unsupported parameter for module: role

Change-Id: Id60e3f06e90e708d9054b905bbfd28ff5450fa49
Closes-Bug: #1691388

The Orchestration service automatically assigns the heat_stack_user
role to users that it creates during stack deployment.
To avoid conflicts, do not add this role to users with the heat_stack_owner role.

Closes-Bug: #1690975

Change-Id: I7a4cb7f4a13de7be4fc9ce9c24057ece6a0ced5c

If enable vpnaas,service_plugins value is 'neutron_vpnaas.services
.vpn.plugin.VPNDriverPlugin'.
It throw warning for 'stevedore.named [-]Could not load
neutron_vpnaas.services.vpn.plugin.VPNDriverPlugin'

Closes-Bug: #1690693

Change-Id: Ia00f733da2dcbdd50e3d62dfe98c8f44cc4a1d26

If enable lbaas,service_plugins value is 'neutron_lbaas.services
.loadbalancer.plugin:LoadBalancerPluginv2'.
It throw warning for 'stevedore.named [-]Could not load neutron_lbaas
.services.loadbalancer.plugin.LoadBalancerPluginv2'

Closes-Bug: #1690694

Change-Id: I38d61405710ca88fbdb3cfd72a534986cf933195

Option "notify_on_state_change" from group "DEFAULT" is deprecated. Use
option "notify_on_state_change" from group “notifications".

refer to
https://docs.openstack.org/ocata/config-reference/compute/config-options.html

Change-Id: I18e80ccd5981b81bfd597b94717ed77b13c05222
Closes-Bug: #1690494

On RedHat libvirt is just a metapackage and it does
nothing with libvirt daemon itself if we remove it.
On Ubuntu the correct package name with libvirt daemon
is libvirt-bin.

Closes-Bug: #1690356
Change-Id: Ief8dfd8c038b10878621aefcc8f559aa20dc498b
Signed-off-by: Pavel Glushchak <pglushchak@virtuozzo.com>

Closes-Bug: #1690302
Change-Id: Ice9e6ece862c4102883c6a5f3bbfe2d291077c8d

when statements should not include jinja2 templating delimiters such as
{{ }} or {% %}. Found: {{ ceph_enable_cache | bool }}
Closes-Bug: #1689550

Change-Id: I2ca039b80260cd753026f8fb077c7d25e46c8b13

During cleanup-host, we need remove swift references from fstab.

Change-Id: I18e9a84b0664eb834cf397bcf6d2406abcab47a7
Closes-Bug: #1689285

The local keyword is used to allow chronyd to appear synchronised to
real time (from the viewpoint of clients polling it), even if it has no
current synchronisation source.

Change-Id: I27d945abb87a37d2fdf972a58c51e2f0994e3de8
Closes-Bug: #1688682

* Use absolute path for panko.conf.j2
* Fix the destination path.

Change-Id: I3abc0bf000be8af3a0cde1f43901245e027bb469
Closes-Bug: #1688681

This patch fix log rotation issue on the following services/program:
- ceph
- chrony
- collectd
- congress
- etcd
- influxdb
- ironic
- kibana
- kuryr
- mongodb
- rally
- tempest
- trove

Change-Id: I1246c071b7340b3267226cefb64a04da1cce9f26
Closes-Bug: #1688649

neutron-openvswitch-agent enter in restarting status in multinode
deployments because br-ex is no longer created on compute nodes
unless DVR is enabled.

This change add a conditional to only add bridge_mappings when
the node is in network or in compute when dvr is enabled.

Change-Id: I9d49158127af7fd78ecbc3d64a7509d1e4ae5672
Closes-Bug: #1688571

Cinder backup requires iSCSI volume and host mount for backing
any iSCSI volume, not only from LVM backend. 'enable_iscsid' is
a catch-all setting for all iSCSI backends that also runs
necessary iSCSI client on the same host.

Change-Id: Ic486a9f72820df2e0a48a18166f043e5ddb5bd0a
Closes-Bug: #1688519

Chrony drops root to user 'chrony' on startup, so it's config file
should be owned by user this user.

Change-Id: I8ba976cc685d2e50e09211f67507487a1e0e3c5f

wait_for module waits 300 seconds for the port started or stopped.  This
is meaningless and useless in precheck. This patch change timeout to 1
seconds.

Change-Id: I9b251ec4ba17ce446655917e8ef5e152ef947298
Closes-Bug: #1688152

Change-Id: I9978cb714e423863afd5b892419f4176b6d73d92
Closes-Bug: #1686761

sahara-engine service need to run ip netns command in container
when set use_namespaces=true in saraha.conf

Change-Id: I8b10cdd0127e1895ae40dfeed1a5910e6bcd701c
Related-Bug: #1687602
Closes-Bug: #1687624

Change-Id: Ieb6ad24bf76efaafa0f8e7391ad671ce094a8ece
Closes-Bug: #1687635

When Kibana interface is opened for the first time, it requires creating
a default index pattern. To view, analyse and search logs, at least one
index pattern has to be created.
This patch automate "flog-*" default index pattern creation and set it
has default config.

Partially-implements: blueprint default-kibana-dashboard
Change-Id: Ie36696f9ad38ba7e49e65e0793a3b98d9f03ee8d

The following Neutron agents:
- neutron-metadata
- neutron-dhcp-agent
- neutron-l3-agent
- neutron-lbaasv2-agent
- neutron-vpn-agent
doesn't require ml2_conf.ini file.
ml2_conf.ini file is used to managed L2 network configuration.
This config doesn't have to be in DHCP, L3, metadata agents...

We should remove it to avoid restarting these agents in case of
ml2_conf.ini file modification.
Only neutron-server, neutron-openvswitch-agent,
neutron-linuxbridge-agent must be restarted.

Closes-Bug: #1677163
Change-Id: I0876b8a3845d1c2bccd996426a65df1a3a6f7085

With nova cells_v2 at least one compute node is required in
inventory to deploy nova.

This change add prechecks to ensure at least 1 compute is present.

Change-Id: I242518ad3bd149ad245515299301777f6b3bdd08
Closes-Bug: #1686410

Telegraf always restart containers when doing reconfigure, due to
"pid_mode" attribut missing during compare_container.
Hence compare_container always evaluated to True.

Change-Id: I335816ebf99a218a5400b2baae5b55788459bfff
Closes-Bug: #1686349

Rally and Tempest containers doesn't required "privileged" attribut.
This must be cleanup.

Change-Id: Ic8de5162523314dd78dad7d25b36145350749bcc

In order to automate Grafana configuration, this fix automatically
set the InfluxDB datasource. This avoid doing a annoying manual
configuration after Kolla-ansible deployment.

Change-Id: I2b1a63fd371966192f1df0a82cee4711c6324710

YAML files are supposed to start with three dashes to indicate the
beginning of the document.
Nota: Ansible won’t complain if they are missing.

Change-Id: I5cd417971f7af2977e7527b7b454a999718cfcda

The pxelinux.cfg directory gets created but isn't owned by the correct
user. This patch ensures that the permissions are correctly updated.

Change-Id: Ifcb80018b72d40c5d4eccf059d1c3442b71be6f8

Change-Id: I6c769e5231e9d0e327ef6b9928666369d694f5fe
Closes-Bug: #1685692

When docker-py is newer than docker daemon following
APIError may be raised:

client is newer than server (client API version: 1.24, server API version: 1.22)

So it's better to pass api_version='auto' to docker client,
so client will discover suitable version if not specified.

Change-Id: Ib747ed91b315ca762936a6efa6b71a29b2d2ae1e
Signed-off-by: Pavel Glushchak <pglushchak@virtuozzo.com>

The OpenStack Integration Test Suite (Tempest) doesn't use oslo.policy
library.
oslo.policy is not used in codebase and not set in requirements.txt:
https://github.com/openstack/tempest/blob/master/requirements.txt#L14

All policy.json related stuff must be removed.

Change-Id: I0d8ee710979a7eebba1e8fc176450e4c0bcbe757

In "Checking free port for Kibana Server" task the storage interface is
used instead of the api interface one.
This patch fix it.

Change-Id: I0b38688aa3774ae4a629cce3161a730fff907a9c

The path of firewall plugin in kolla-ansible is wrong.

Change-Id: I979af769d8bc31e3b6373a629e9e090b987d0058
Closes-Bug: #1684459

When using cinder lvm backend storage driver, lvcreate will load dm_*
linux modules when need. So /lib/modules is needed by cinder_volume
container.

Change-Id: I80985794d6aaa6b4ddc71680a97e0d07085dafa7
Closes-Bug: #1684378

Currently Cinder configuration hard-codes tgtadm as the only
iSCSI management utility. tgtd and tgtadm are older tools to
manage iSCSI volumes. Newer versions of Linux distros use
targetcli and kernel based iSCSI facilities (LIO) and can be
a preferred way for some users.

Change-Id: I10c0a7d596994a9201218e58ca2ad184774a8ce2
Closes-Bug: #1684139

When using panko event dispatcher and publisher in ceilometer, it
depends on panko.conf file.

Change-Id: Ie91c072b233597758955b70bd526b2603b86e995
Closes-Bug: #1672241

In task "Running Neutron vpnaas bootstrap container",
parameter "volumes" of kolla_docker should be
"{{ neutron_vpnaas_agent.volumes }}"
NOT
"{{ neutron_vpnaas_agent.image }}"

Change-Id: Id7c95d4d6f53dfab384e387951eecf4d87459de
Closes-Bug: #1683669

Change-Id: I3b3ad184dea6b93f21c8778386c04a93b38bc2af
Closes-Bug: #1683802

Simple_cell_setup is not recomended to use.
Is better create map_cell0 manually, create base
cell for non cell deployments and run discover_hosts.

This PS migrate actual config to make use of described
workflow at [1]. We our actual workflow we're running
into the issue that services are not mapped until cells
are present, breaking deployment waiting for compute
services to appear.

[1] https://docs.openstack.org/developer/nova/cells.html#fresh-install

Change-Id: Id061e8039e72de77a04c51657705457193da2d0f
Closes-Bug: #1682060

"/lib/modules" volume mount has been introduced for the following
reason:
- "ebtables" in case of neutron-linuxbridge-agent container
- "ipsec" in case of neutron-vpnaas-agent container
- "openvswitch" in case of openvswitch-vswitchd container

For all other containers this is not required and must be removed.

Change-Id: I11610c060f05d2c993394adabd15aa080de3bc24