apex-tripleo-heat-templates.git
7 years agoscenario002/container: run Barbican non-containerized
Emilien Macchi [Tue, 15 Aug 2017 03:46:10 +0000 (20:46 -0700)]
scenario002/container: run Barbican non-containerized

... until https://review.openstack.org/#/c/474327 is merged.
In the meantime, let's test the scenario with Barbican like before.

Depends-On: Ib5c99482f62397fc5fb79a9dc537dfb06ee7f4df
Change-Id: Ia96736ad3ddabd33c5ee4518a3f63bafeffcf391

7 years agoMerge "Set default OSD pool size to 1 in scenario 001/004 containers"
Jenkins [Thu, 17 Aug 2017 04:37:16 +0000 (04:37 +0000)]
Merge "Set default OSD pool size to 1 in scenario 001/004 containers"

7 years agoMerge "Containerize virtlogd"
Jenkins [Thu, 17 Aug 2017 01:02:22 +0000 (01:02 +0000)]
Merge "Containerize virtlogd"

7 years agoMerge "Delete docker-centos-tripleoupstream.yaml"
Jenkins [Thu, 17 Aug 2017 01:00:06 +0000 (01:00 +0000)]
Merge "Delete docker-centos-tripleoupstream.yaml"

7 years agoMerge "Add ServiceData to hidden params"
Jenkins [Wed, 16 Aug 2017 21:11:48 +0000 (21:11 +0000)]
Merge "Add ServiceData to hidden params"

7 years agoMerge "Add NeutronOverlayIPVersion parameter to neutron-plugins-ml2 service"
Jenkins [Wed, 16 Aug 2017 19:46:32 +0000 (19:46 +0000)]
Merge "Add NeutronOverlayIPVersion parameter to neutron-plugins-ml2 service"

7 years agoMerge "Render VIPs dynamically based on network_data.yaml"
Jenkins [Wed, 16 Aug 2017 15:30:29 +0000 (15:30 +0000)]
Merge "Render VIPs dynamically based on network_data.yaml"

7 years agoSet default OSD pool size to 1 in scenario 001/004 containers
Giulio Fidente [Wed, 16 Aug 2017 11:41:28 +0000 (13:41 +0200)]
Set default OSD pool size to 1 in scenario 001/004 containers

When the OSD pool size is unset it defaults to 3, while we only
have a single OSD in CI so the pools are created but not writable.

We did set the default pool size to 1 in the non-containerized
scenarios but apparently missed it in the containerized version.

Change-Id: I1ac1fe5c2effd72a2385ab43d27abafba5c45d4d
Closes-Bug: #1710773

7 years agoMerge "Don't unregister on system/resource delete"
Jenkins [Wed, 16 Aug 2017 06:43:02 +0000 (06:43 +0000)]
Merge "Don't unregister on system/resource delete"

7 years agoMerge "Bind mount tripleo.cnf in transient bootstrap containers"
Jenkins [Wed, 16 Aug 2017 01:38:32 +0000 (01:38 +0000)]
Merge "Bind mount tripleo.cnf in transient bootstrap containers"

7 years agoMerge "Convert network templates to be rendered via j2"
Jenkins [Wed, 16 Aug 2017 01:38:25 +0000 (01:38 +0000)]
Merge "Convert network templates to be rendered via j2"

7 years agoMerge "Consolidate deployment in major-upgrade-composable-steps"
Jenkins [Wed, 16 Aug 2017 00:39:01 +0000 (00:39 +0000)]
Merge "Consolidate deployment in major-upgrade-composable-steps"

7 years agoMerge "Do not run clustercheck on the host after O->P upgrade"
Jenkins [Tue, 15 Aug 2017 23:53:57 +0000 (23:53 +0000)]
Merge "Do not run clustercheck on the host after O->P upgrade"

7 years agoMerge "Internal TLS support for mongodb container"
Jenkins [Tue, 15 Aug 2017 23:53:29 +0000 (23:53 +0000)]
Merge "Internal TLS support for mongodb container"

7 years agoMerge "Set file mode permission of Ceph keyrings"
Jenkins [Tue, 15 Aug 2017 23:53:22 +0000 (23:53 +0000)]
Merge "Set file mode permission of Ceph keyrings"

7 years agoMerge "Make network-isolation-v6 environment rendered for all roles"
Jenkins [Tue, 15 Aug 2017 23:08:42 +0000 (23:08 +0000)]
Merge "Make network-isolation-v6 environment rendered for all roles"

7 years agoMerge "Fix parsing of DockerCephDaemonImage parameter"
Jenkins [Tue, 15 Aug 2017 16:12:19 +0000 (16:12 +0000)]
Merge "Fix parsing of DockerCephDaemonImage parameter"

7 years agoAdd NeutronOverlayIPVersion parameter to neutron-plugins-ml2 service
Feng Pan [Tue, 16 May 2017 19:22:47 +0000 (15:22 -0400)]
Add NeutronOverlayIPVersion parameter to neutron-plugins-ml2 service

This patch adds NeutronOverlayIPVersion parameter to congfigure
neutron ML2 overlay_ip_version option from T-H-T. puppet-neutron
already has support for configuration of this option, we are just
exposing it from T-H-T. This parameter needs to be set to '6' when
IPv6 vxlan tunnel endpoints are desired.

Closes-Bug: #1691213

Change-Id: I056afa25f67a3b6857bdfef14e6d582b0a9e5e93
Signed-off-by: Feng Pan <fpan@redhat.com>
7 years agoMerge "Fix Heat condition for RHEL registration yum update"
Jenkins [Tue, 15 Aug 2017 14:34:38 +0000 (14:34 +0000)]
Merge "Fix Heat condition for RHEL registration yum update"

7 years agoConvert network templates to be rendered via j2
Steven Hardy [Wed, 9 Aug 2017 18:17:49 +0000 (19:17 +0100)]
Convert network templates to be rendered via j2

Use the network.network.j2.yaml to render these files, instead
of relying on the hard-coded versions.

Note this doesn't currently consider the _v6 templates as we may want
to deprecate these and instead rely on an ipv6 specific network_data file,
or perhaps make the network/network.network.j2.yaml generic and able to
detect the version from the cidr?

Change-Id: I662e8d0b3737c7807d18c8917bfce1e25baa3d8a
Partially-Implements: blueprint composable-networks

7 years agoMerge "Enable TLS configuration for containerized Galera"
Jenkins [Mon, 14 Aug 2017 23:03:35 +0000 (23:03 +0000)]
Merge "Enable TLS configuration for containerized Galera"

7 years agoMerge "Make HA container bundle work on remote nodes"
Jenkins [Mon, 14 Aug 2017 23:03:27 +0000 (23:03 +0000)]
Merge "Make HA container bundle work on remote nodes"

7 years agoMerge "Convert cephstorage-role.yaml to role.role.j2.yaml"
Jenkins [Mon, 14 Aug 2017 23:03:17 +0000 (23:03 +0000)]
Merge "Convert cephstorage-role.yaml to role.role.j2.yaml"

7 years agoMerge "Fix metadata_settings in containerized mongodb"
Jenkins [Mon, 14 Aug 2017 22:52:21 +0000 (22:52 +0000)]
Merge "Fix metadata_settings in containerized mongodb"

7 years agoDelete docker-centos-tripleoupstream.yaml
Steve Baker [Wed, 26 Jul 2017 22:23:39 +0000 (10:23 +1200)]
Delete docker-centos-tripleoupstream.yaml

This file is generated and needs to be manually maintained. It
would be better for users who want to deploy latest directly from
docker hub to generate it locally by running:

    openstack overcloud container image prepare \
      --namespace tripleoupstream \
      --tag latest \
      --env-file docker-centos-tripleoupstream.yaml

The documentation and CI are being updated to use prepare.

Change-Id: I86503f1076459ae9d84a34e649a6097cba10fa3c
Closes-Bug: #1696598

7 years agoMerge "Enable TLS for nova api and placement containers"
Jenkins [Mon, 14 Aug 2017 22:02:25 +0000 (22:02 +0000)]
Merge "Enable TLS for nova api and placement containers"

7 years agoMerge "Make containerized nova-api run with httpd"
Jenkins [Mon, 14 Aug 2017 22:01:04 +0000 (22:01 +0000)]
Merge "Make containerized nova-api run with httpd"

7 years agoSet file mode permission of Ceph keyrings
John Fulton [Wed, 9 Aug 2017 21:07:40 +0000 (17:07 -0400)]
Set file mode permission of Ceph keyrings

Pass mode parameter to ceph-ansible in place of ACLs parameter
because ACLs are not for same UID in container as container host
and because ACLs are not passed by kolla_config.

Change-Id: I7e3433eab8e2a62963b623531f223d5abd301d16
Closes-Bug: #1709683

7 years agoDon't unregister on system/resource delete
James Slagle [Fri, 11 Aug 2017 12:05:10 +0000 (08:05 -0400)]
Don't unregister on system/resource delete

Don't unregister systems from the portal/satellite
when deleting from Heat. There are several reasons why
it's compelling to fix this behavior. See
https://bugs.launchpad.net/tripleo/+bug/1710144
for full information. The previous behavior can be triggered
by setting the DeleteOnRHELUnregistration parameter to "true".

Closes-Bug: #1710144
Change-Id: I909a6f7a049dc23fc27f2231a4893d428f06a1f1

7 years agoFix Heat condition for RHEL registration yum update
James Slagle [Thu, 10 Aug 2017 17:42:31 +0000 (13:42 -0400)]
Fix Heat condition for RHEL registration yum update

There were 2 problems with this condition making the
rhel-registration.yal template broken:

"conditions" should be "condition"

The condition should refer to just a condition name defined in the
"conditions:" section of the template.

Change-Id: I14d5c72cf86423808e81f1d8406098d5fd635e66
Closes-Bug: #1709916

7 years agoFix metadata_settings in containerized mongodb
Damien Ciabrini [Wed, 9 Aug 2017 07:25:42 +0000 (07:25 +0000)]
Fix metadata_settings in containerized mongodb

The containerized version of the mongodb service omits the
metadata_settings definition [1], which confuses certmonger when
internal TLS is enabled and make the generation of certificates fail.

Use the right setting from the non-containerized profile.

[1] https://review.openstack.org/#/c/461780/

Change-Id: I50a9a3a822ba5ef5d2657a12c359b51b7a3a42f2
Closes-Bug: #1709553

7 years agoBind mount tripleo.cnf in transient bootstrap containers
Damien Ciabrini [Fri, 11 Aug 2017 11:24:12 +0000 (11:24 +0000)]
Bind mount tripleo.cnf in transient bootstrap containers

Various containerized services (e.g. nova, neutron, heat) run initial set up
steps with some ephemeral containers that don't use kolla_start. The
tripleo.cnf file is not copied in /etc/my.cnf.d and this can break some
deployments (e.g. when using internal TLS, service lack SSL settings).

Fix the configuration of transient containers by bind mounting of the
tripleo.cnf file when kolla_start is not used.

Change-Id: I5246f9d52fcf8c8af81de7a0dd8281169c971577
Closes-Bug: #1710127
Co-Authored-By: Juan Antonio Osorio Robles <jaosorior@redhat.com>
7 years agoContainerize virtlogd
Jiri Stransky [Tue, 30 May 2017 13:49:51 +0000 (15:49 +0200)]
Containerize virtlogd

So far we've been using virtlogd running on the host, we should now be
using virtlogd from a container.

Co-Authored-By: Martin André <m.andre@redhat.com>
Co-Authored-By: Jiri Stransky <jistr@redhat.com>
Change-Id: I998c69ea1f7480ebb90afb44d6006953a84a1c04

7 years agoConsolidate deployment in major-upgrade-composable-steps
Jose Luis Franco Arza [Mon, 14 Aug 2017 09:49:00 +0000 (11:49 +0200)]
Consolidate deployment in major-upgrade-composable-steps

After 483293 commit is merged, major-upgrade-composable-steps.yaml file
is pointing to the wrong location deployment, which is now under
common/ folder.

Change-Id: Ic6784533d1c21b5b8fcb422bccd820af72e499d9

7 years agoMerge "Pass monitor_address_block to ceph-ansible for mon_host"
Jenkins [Sun, 13 Aug 2017 22:26:57 +0000 (22:26 +0000)]
Merge "Pass monitor_address_block to ceph-ansible for mon_host"

7 years agoMerge "Add environment to disable deploy steps"
Jenkins [Sun, 13 Aug 2017 06:56:00 +0000 (06:56 +0000)]
Merge "Add environment to disable deploy steps"

7 years agoMerge "Add support for update_tasks"
Jenkins [Sun, 13 Aug 2017 06:51:54 +0000 (06:51 +0000)]
Merge "Add support for update_tasks"

7 years agoMerge "Add RoleConfig output"
Jenkins [Sun, 13 Aug 2017 06:51:35 +0000 (06:51 +0000)]
Merge "Add RoleConfig output"

7 years agoMerge "Default docker_puppet_debug to false"
Jenkins [Sun, 13 Aug 2017 06:50:34 +0000 (06:50 +0000)]
Merge "Default docker_puppet_debug to false"

7 years agoMerge "Move deploy-steps-playbook to deploy-steps-tasks"
Jenkins [Sun, 13 Aug 2017 06:45:37 +0000 (06:45 +0000)]
Merge "Move deploy-steps-playbook to deploy-steps-tasks"

7 years agoMerge "Convert blockstorage-role.yaml to role.role.j2.yaml"
Jenkins [Sat, 12 Aug 2017 13:48:35 +0000 (13:48 +0000)]
Merge "Convert blockstorage-role.yaml to role.role.j2.yaml"

7 years agoMerge "Convert objectstorage-role.yaml to role.role.j2.yaml"
Jenkins [Sat, 12 Aug 2017 13:22:39 +0000 (13:22 +0000)]
Merge "Convert objectstorage-role.yaml to role.role.j2.yaml"

7 years agoFix parsing of DockerCephDaemonImage parameter
Giulio Fidente [Tue, 8 Aug 2017 11:25:27 +0000 (13:25 +0200)]
Fix parsing of DockerCephDaemonImage parameter

Splitting by colon using native str_split function did not work well
because we needed a right split.

This change replaces the str_split calls with yaql rightSplit().

Change-Id: Iab2f69a5fadc6b02e2eacf3c9d1a9024b0212ac6

7 years agoPass monitor_address_block to ceph-ansible for mon_host
Giulio Fidente [Tue, 8 Aug 2017 21:18:23 +0000 (23:18 +0200)]
Pass monitor_address_block to ceph-ansible for mon_host

The ip address which clients and other nodes use to connect to the
monitors is derived from the monitor_interface parameter unless
a monitor_address or monitor_address_block is given (to set mon_host
into ceph.conf); this change adds setting for monitor_address_block to
match the public_network so that clients attempt to connect to the mons
on the appropriate network.

Change-Id: I7187e739e9f777eab724fbc09e8b2c8ddedc552d
Closes-Bug: #1709485

7 years agoAdd environment to disable deploy steps
Steven Hardy [Fri, 21 Jul 2017 16:45:09 +0000 (17:45 +0100)]
Add environment to disable deploy steps

This enables either deploying without configuring any services, or
temporarily disabling the deploy steps such as will be required
for minor updates where we want to re-run the rolling update outside
of heat.

To deploy directly via ansible-playbook you can do e.g:

openstack overcloud config download --config-dir tmpconfig
cd tmpconfig/tripleo-6b02U7-config
ansible-playbook -vvv -b -i /usr/bin/tripleo-ansible-inventory deploy_steps_playbook.yaml

Which will run the same ansible steps as we normally run via heat.

Change-Id: I59947b67523dfcc43d454d4ac7d82b06804cf71d

7 years agoAdd support for update_tasks
Steven Hardy [Fri, 21 Jul 2017 10:43:25 +0000 (11:43 +0100)]
Add support for update_tasks

These work the same way as upgrade_tasks *but* they use a step variable
instead of tags, so we can iterate over a count/sequence which isn't
possibly via a wrapper playbook with tags (we may want to align upgrade
tasks with the same approach if this works out well).

Note the tasks can be run via ansible-playbook on the undercloud, like:

openstack overcloud config download --config-dir tmpconfig
cd tmpconfig/tripleo-HCrDA6-config
ansible-playbook -b -i /usr/bin/tripleo-ansible-inventory update_steps_playbook.yaml --limit controller

The above will do a rolling update for the Controller role (note the inconsistent
capitalization, we probably need to fix the group naming in tripleo-ansible-inventory)
because we specify serial: 1 in the playbook.

You can also trigger an update explicitly on one node like this, which is useful for debugging:

ansible-playbook -vvv -b -i /usr/bin/tripleo-ansible-inventory update_steps_playbook.yaml --limit overcloud-controller-0

Change-Id: I20bb3e26ab9d9cadf1a31fd304de8a014a901aa9

7 years agoAdd RoleConfig output
Steven Hardy [Thu, 20 Jul 2017 16:11:44 +0000 (17:11 +0100)]
Add RoleConfig output

This exposes the deploy workflow for all roles from deploy-steps
via overcloud.j2.yaml - which means we can write it via the new
openstack overcloud config download command and/or run the workflow
outside of heat via mistral

With https://review.openstack.org/#/c/485732/ applied to
tripleoclient it becomes possible to do:

openstack overcloud config download --config-dir tmpconfig
cd tmpconfig/tripleo-EvEZk0-config
ansible-playbook -b -i /usr/bin/tripleo-ansible-inventory deploy_steps_playbook.yaml

This runs the deploy steps, exactly the same as normally run via heat
via ansible-playbook for all overcloud nodes (--limit can be used to restrict
to specific nodes/roles).

Change-Id: I96ec09bc788836584c4b39dcce5bf9b80e914c71

7 years agoDefault docker_puppet_debug to false
Steven Hardy [Fri, 21 Jul 2017 10:21:09 +0000 (11:21 +0100)]
Default docker_puppet_debug to false

This isn't set unless the playbook is run via heat, so default it to false
to enable easier use via ansible-playbook combined with tripleo-ansible-inventory

Change-Id: I9705e4533831a019dd0051e5522d4b7958682506

7 years agoMove deploy-steps-playbook to deploy-steps-tasks
Steven Hardy [Thu, 20 Jul 2017 14:04:10 +0000 (15:04 +0100)]
Move deploy-steps-playbook to deploy-steps-tasks

So that we can more easily iterate over an include in an output

Change-Id: Idd5bb47589e5c37123caafcded1afbff8881aa33

7 years agoMerge "Consolidate puppet/docker deployments with one deploy steps workflow"
Jenkins [Sat, 12 Aug 2017 07:18:07 +0000 (07:18 +0000)]
Merge "Consolidate puppet/docker deployments with one deploy steps workflow"

7 years agoMerge "Correct gnocchi-upgrade command quotes"
Jenkins [Sat, 12 Aug 2017 04:44:00 +0000 (04:44 +0000)]
Merge "Correct gnocchi-upgrade command quotes"

7 years agoMerge "Convert compute-role.yaml to role.role.j2.yaml"
Jenkins [Sat, 12 Aug 2017 03:19:08 +0000 (03:19 +0000)]
Merge "Convert compute-role.yaml to role.role.j2.yaml"

7 years agoMerge "Convert controller-role.yaml to role.role.j2.yaml"
Jenkins [Sat, 12 Aug 2017 03:17:49 +0000 (03:17 +0000)]
Merge "Convert controller-role.yaml to role.role.j2.yaml"

7 years agoAdd ServiceData to hidden params
Ben Nemec [Fri, 11 Aug 2017 21:16:24 +0000 (16:16 -0500)]
Add ServiceData to hidden params

It looks like this was added recently and it doesn't appear to be a
parameter we want in the sample environments.

Change-Id: I0ac433553e7ad9b0a54c011b66c54b4692b44be0

7 years agoMerge "TLS everywhere: Configure CA for mongodb"
Jenkins [Fri, 11 Aug 2017 21:20:59 +0000 (21:20 +0000)]
Merge "TLS everywhere: Configure CA for mongodb"

7 years agoMerge "Add script to create tripleo-admin on deployed servers"
Jenkins [Fri, 11 Aug 2017 19:07:26 +0000 (19:07 +0000)]
Merge "Add script to create tripleo-admin on deployed servers"

7 years agoCorrect gnocchi-upgrade command quotes
Jose Luis Franco Arza [Tue, 8 Aug 2017 14:19:20 +0000 (16:19 +0200)]
Correct gnocchi-upgrade command quotes

After merging commit 488796, single quotation marks
were missed. This causes the upgrade to fail as the
flag --sacks-number is considered a su command flag.

Also mounts Ceph config data into the container which
seems needed for the gnocchi-upgrade command when
configured to use Ceph.

Also move the gnocchi db sync to step 4, so ceph is
ready. Add a retry loop to ceilometer-upgrade cmd so
it doesnt fail while apache is restarted.

Closes-Bug: #1709322
Change-Id: I62f3a5fa2d43a2cd579f72286661d503e9f08b90

7 years agoMerge "openstack-heat-templates: fix deprecation path"
Jenkins [Fri, 11 Aug 2017 17:55:43 +0000 (17:55 +0000)]
Merge "openstack-heat-templates: fix deprecation path"

7 years agoConsolidate puppet/docker deployments with one deploy steps workflow
Steven Hardy [Thu, 13 Jul 2017 12:40:48 +0000 (13:40 +0100)]
Consolidate puppet/docker deployments with one deploy steps workflow

If we consolidate these we can focus on one implementation (the new ansible
based one used for docker-steps)

Change-Id: Iec0ad2278d62040bf03613fc9556b199c6a80546
Depends-On: Ifa2afa915e0fee368fb2506c02de75bf5efe82d5

7 years agoConvert cephstorage-role.yaml to role.role.j2.yaml
Steven Hardy [Tue, 11 Jul 2017 16:28:38 +0000 (17:28 +0100)]
Convert cephstorage-role.yaml to role.role.j2.yaml

Add some special-casing for backwards compatibility, such that the
CephStorage role can be rendered via j2 for support of composable networks.

Change-Id: Iee92bb6ee94963717d3a8ef400e7970f62576a0d
Partially-Implements: blueprint composable-networks

7 years agoConvert blockstorage-role.yaml to role.role.j2.yaml
Steven Hardy [Tue, 11 Jul 2017 16:22:28 +0000 (17:22 +0100)]
Convert blockstorage-role.yaml to role.role.j2.yaml

Add some special-casing for backwards compatibility, such that the
BlockStorage role can be rendered via j2 for support of composable networks.

Change-Id: Ia5fb5ff6dbe218710e95a69583ac289cf7b4af9e
Partially-Implements: blueprint composable-networks

7 years agoConvert objectstorage-role.yaml to role.role.j2.yaml
Steven Hardy [Tue, 11 Jul 2017 16:09:04 +0000 (17:09 +0100)]
Convert objectstorage-role.yaml to role.role.j2.yaml

Add some special-casing for backwards compatibility, such that the
ObjectStorage role can be rendered via j2 for support of composable networks.

Change-Id: I52abbefe2f5035059ccbed925990faab020c6c89
Partially-Implements: blueprint composable-networks

7 years agoConvert compute-role.yaml to role.role.j2.yaml
Steven Hardy [Tue, 11 Jul 2017 14:45:52 +0000 (15:45 +0100)]
Convert compute-role.yaml to role.role.j2.yaml

Add some special-casing for backwards compatibility, such that the
Compute role can be rendered via j2 for support of composable networks.

Change-Id: Ieee446583f77bb9423609d444c576788cf930121
Partially-Implements: blueprint composable-networks

7 years agoConvert controller-role.yaml to role.role.j2.yaml
Steven Hardy [Tue, 4 Jul 2017 17:20:10 +0000 (18:20 +0100)]
Convert controller-role.yaml to role.role.j2.yaml

Add deprecated role-specific parameters to role definition, in
order to special-case some parameters for backwards compatibility,
such that the Controller role can be rendered via j2 for support
of composable networks.

Co-Authored By: Dan Sneddon <dsneddon@redhat.com>
Change-Id: I5983f03ae1b7f0b6add793914540b8ca405f9b2b
Partially-Implements: blueprint composable-networks

7 years agoInternal TLS support for mongodb container
Juan Antonio Osorio Robles [Fri, 11 Aug 2017 08:46:49 +0000 (11:46 +0300)]
Internal TLS support for mongodb container

This bind mounts the necessary files for the mongodb container to serve
TLS in the internal network.

bp tls-via-certmonger-containers

Change-Id: Ieef2a456a397f7d5df368ddd5003273cb0bb7259
Co-Authored-By: Damien Ciabrini <dciabrin@redhat.com>
7 years agoTLS everywhere: Configure CA for mongodb
Juan Antonio Osorio Robles [Fri, 11 Aug 2017 13:07:13 +0000 (16:07 +0300)]
TLS everywhere: Configure CA for mongodb

It wasn't being configured, thus making mongodb fail.

Change-Id: If0d7513aacfa74493a9747440fb97f915a77db84
Closes-Bug: #1710162

7 years agoMerge "Move HAProxy's public TLS logic from controller to service template"
Jenkins [Fri, 11 Aug 2017 12:02:30 +0000 (12:02 +0000)]
Merge "Move HAProxy's public TLS logic from controller to service template"

7 years agoMerge "Set virsh secret with an init step when using Ceph"
Jenkins [Fri, 11 Aug 2017 11:28:41 +0000 (11:28 +0000)]
Merge "Set virsh secret with an init step when using Ceph"

7 years agoMerge "Keep dynamic network creation backward compatible."
Jenkins [Fri, 11 Aug 2017 11:26:41 +0000 (11:26 +0000)]
Merge "Keep dynamic network creation backward compatible."

7 years agoEnable TLS for nova api and placement containers
Juan Antonio Osorio Robles [Wed, 2 Aug 2017 06:58:46 +0000 (09:58 +0300)]
Enable TLS for nova api and placement containers

With these two services running over httpd in the containers, we can now
enable TLS for them.

bp tls-via-certmonger-containers

Change-Id: Ib8fc37a391e3b32feef0ac6492492c0088866d21

7 years agoMake containerized nova-api run with httpd
Juan Antonio Osorio Robles [Mon, 19 Jun 2017 12:21:15 +0000 (15:21 +0300)]
Make containerized nova-api run with httpd

The non-containerized version will run over httpd [1], and for the
containerized TLS work, it is needed in the container version as well.

[1] Iac35b7ddcd8a800901548c75ca8d5083ad17e4d3

bp tls-via-certmonger-containers

Depends-On: I1c5f13039414f17312f91a5e0fd02019aa08e00e
Change-Id: I2c39a2957fd95dd261b5b8c4df5e66e00a68d2f7

7 years agoEnable TLS configuration for containerized Galera
Damien Ciabrini [Wed, 2 Aug 2017 10:13:48 +0000 (06:13 -0400)]
Enable TLS configuration for containerized Galera

In non-containerized deployments, Galera can be configured to use TLS
for gcomm group communication when enable_internal_tls is set to true.

Fix the metadata service definition and update the Kolla configuration
to make gcomm use TLS in containers, if configured.

bp tls-via-certmonger-containers

Change-Id: Ibead27be81910f946d64b8e5421bcc41210d7430
Co-Authored-By: Juan Antonio Osorio Robles <jaosorior@redhat.com>
Closes-Bug: #1708135
Depends-On: If845baa7b0a437c28148c817b7f94d540ca15814

7 years agoMove HAProxy's public TLS logic from controller to service template
Juan Antonio Osorio Robles [Mon, 7 Aug 2017 11:25:38 +0000 (14:25 +0300)]
Move HAProxy's public TLS logic from controller to service template

This de-couples public TLS from controllers to now run wherever HAProxy
is deployed.

Partially-Implements: blueprint composable-networks
Change-Id: I9e84a25a363899acf103015527787bdd8248949f

7 years agoMerge "Noop controller pre and post config resources."
Jenkins [Thu, 10 Aug 2017 21:49:55 +0000 (21:49 +0000)]
Merge "Noop controller pre and post config resources."

7 years agoMerge "Fix cidr get_attr in custom networks"
Jenkins [Thu, 10 Aug 2017 19:41:17 +0000 (19:41 +0000)]
Merge "Fix cidr get_attr in custom networks"

7 years agoMerge "Create parameters for haproxy TLS certs and keys"
Jenkins [Thu, 10 Aug 2017 19:37:55 +0000 (19:37 +0000)]
Merge "Create parameters for haproxy TLS certs and keys"

7 years agoMerge "Docker/TLS everywhere: Add telemetry and neutron services to environment"
Jenkins [Thu, 10 Aug 2017 04:32:32 +0000 (04:32 +0000)]
Merge "Docker/TLS everywhere: Add telemetry and neutron services to environment"

7 years agoMerge "Addition of Nuage as mechanism driver for ML2"
Jenkins [Wed, 9 Aug 2017 15:11:25 +0000 (15:11 +0000)]
Merge "Addition of Nuage as mechanism driver for ML2"

7 years agoSet virsh secret with an init step when using Ceph
Giulio Fidente [Wed, 9 Aug 2017 10:13:46 +0000 (12:13 +0200)]
Set virsh secret with an init step when using Ceph

Run virsh secret-define and secret-set-value in an init step
instead of relying on the puppet-nova exec.

Co-Authored-By: Jiri Stransky <jistr@redhat.com>
Change-Id: Ic950e290af1c66d34b40791defbdf4f8afaa11da
Closes-Bug: #1709583

7 years agoMake network-isolation-v6 environment rendered for all roles
Michele Baldessari [Thu, 15 Jun 2017 08:22:21 +0000 (10:22 +0200)]
Make network-isolation-v6 environment rendered for all roles

In change If3989f24f077738845d2edbee405bd9198e7b7db we moved to jinja2
templating to render the networks. This change aims at doing so for the
IPv6 network isolation environment.

Change-Id: Ieebcff3db3f5756a5d23080ea3d09ce78de69e21

7 years agoMerge "Use number for KeystoneCronTokenFlushMaxDelay instead of string"
Jenkins [Wed, 9 Aug 2017 07:42:18 +0000 (07:42 +0000)]
Merge "Use number for KeystoneCronTokenFlushMaxDelay instead of string"

7 years agoMerge "Don't curl metadata server in userdata example"
Jenkins [Wed, 9 Aug 2017 04:41:12 +0000 (04:41 +0000)]
Merge "Don't curl metadata server in userdata example"

7 years agoMerge "MariaDB: create clustercheck user at container bootstrap"
Jenkins [Tue, 8 Aug 2017 21:53:34 +0000 (21:53 +0000)]
Merge "MariaDB: create clustercheck user at container bootstrap"

7 years agoMake HA container bundle work on remote nodes
Michele Baldessari [Tue, 8 Aug 2017 19:27:48 +0000 (21:27 +0200)]
Make HA container bundle work on remote nodes

Right now when we deploy an HA bundle on a pacemaker remote node,
the deploy will fail due to the fact that the bundle includes
tripleo::profile::base::pacemaker which makes a call to
hiera('hacluster_pwd') which will fail on pcmk remote nodes.
While we could noop the profile on pcmk nodes, it's much simpler
to just make sure this hiera key exists on pcmk remote nodes.

Also make sure that pacemaker::corosync::manage_fw is set to false
on remote nodes, otherwise the mere inclusion of the pacemaker
profile will cause iptables-save to run in a container and thus failing.

Change-Id: I09b3e54a470cc2d600a701d23463962501c5c9d6

7 years agoFix cidr get_attr in custom networks
Giulio Fidente [Tue, 8 Aug 2017 19:00:30 +0000 (21:00 +0200)]
Fix cidr get_attr in custom networks

We were missing the square brackets around the list of arguments
for get_attr when building the networks cidr output.

This passed CI because Heat does not fail validation and Ceph (which
is consuming the cidr output) is tested with a single network (ctlplane)
which does not build the output using the same templates.

Change-Id: I40bba0784a30295cb0d4eda1fbff20ebac85db99
Closes-Bug: #1709464

7 years agoDocker/TLS everywhere: Add telemetry and neutron services to environment
Juan Antonio Osorio Robles [Tue, 8 Aug 2017 12:35:05 +0000 (12:35 +0000)]
Docker/TLS everywhere: Add telemetry and neutron services to environment

some resources were missing, so this syncs up what's working right now.

bp tls-via-certmonger-containers

Change-Id: Ic8fe20d0240f1ad8f18218d66634029d522d4d5a

7 years agoKeep dynamic network creation backward compatible.
Sofer Athlan-Guyot [Mon, 7 Aug 2017 14:04:08 +0000 (16:04 +0200)]
Keep dynamic network creation backward compatible.

We had an history mapping for InternalApi to InternalNetwork.  If we
remove it then heat will want to destroy InternalNetwork and create
InternalApi which cannot work during upgrade.

This adds compat name parameters to network_data.yaml.

Closes-Bug: #1709105

Change-Id: I8ce6419a5e13a13ee6e991db5ca2196763f52d7a

7 years agoAdd script to create tripleo-admin on deployed servers
Jiri Stransky [Thu, 3 Aug 2017 12:23:27 +0000 (14:23 +0200)]
Add script to create tripleo-admin on deployed servers

When using deployed servers, we want to create a standard
tripleo-admin user for Mistral's ssh tasks (e.g. running Ansible on
overcloud). This script wraps the respective Mistral workflow.

Change-Id: I2de698b4aae07f74569243a9e7c1c56eb578e700
Related-Bug: #1708180
Depends-On: Ibe8e54f7b38d8c6c8d944d2b13f0eed004c34c4c

7 years agoCreate parameters for haproxy TLS certs and keys
Juan Antonio Osorio Robles [Mon, 7 Aug 2017 08:01:24 +0000 (11:01 +0300)]
Create parameters for haproxy TLS certs and keys

this removes the hardcoded paths for the haproxy certs and keys and will
enable re-use. We'll use this in a further commit in the containterized
TLS work.

Change-Id: I602e5a569e2e7e60835deb80532abcedd7a1f63d

7 years agoUse number for KeystoneCronTokenFlushMaxDelay instead of string
Juan Antonio Osorio Robles [Fri, 4 Aug 2017 05:36:42 +0000 (08:36 +0300)]
Use number for KeystoneCronTokenFlushMaxDelay instead of string

Using a string results in an erroneous check in puppet-keystone, which
sets up a zero where it shouldn't. So we change it to number to avoid
that. Note that there will also be a puppet-keystone fix for this.
Changing the value here assures that deployers only give valid values to
this parameter.

Change-Id: I00823e23358df91ce54f421c12636f05d4196e15
Closes-Bug: #1708584

7 years agoMerge "Add Telemetry services to scenario002"
Jenkins [Sat, 5 Aug 2017 16:22:56 +0000 (16:22 +0000)]
Merge "Add Telemetry services to scenario002"

7 years agoMerge "Start redis service after upgrade"
Jenkins [Sat, 5 Aug 2017 15:22:56 +0000 (15:22 +0000)]
Merge "Start redis service after upgrade"

7 years agoMerge "Stop and disable openstack-nova-compute service on compute nodes"
Jenkins [Fri, 4 Aug 2017 20:47:16 +0000 (20:47 +0000)]
Merge "Stop and disable openstack-nova-compute service on compute nodes"

7 years agoMerge "Run gnocchi upgrade with sacks in docker template"
Jenkins [Fri, 4 Aug 2017 20:47:07 +0000 (20:47 +0000)]
Merge "Run gnocchi upgrade with sacks in docker template"

7 years agoMerge "Change the directory for haproxy certs/keys to be service-specific"
Jenkins [Fri, 4 Aug 2017 14:04:44 +0000 (14:04 +0000)]
Merge "Change the directory for haproxy certs/keys to be service-specific"

7 years agoRender VIPs dynamically based on network_data.yaml
Dan Sneddon [Fri, 21 Jul 2017 21:33:33 +0000 (14:33 -0700)]
Render VIPs dynamically based on network_data.yaml

This change modifies the templates to dynamically define the VIPs
based on network_data.yaml. If a network is defined and marked
with "vip: true" in network_data.yaml, it will be included in the
overcloud.yaml which defines the deployment-level resources.

This should make it possible to create custom networks and
use them for services which use high-availability through VIPs.

Also, extraconfig/nova_metadata/krb-service-pricipals.yaml
was modified to dynamically produce the FQDN map for VIPs on
isolated networks, to match overcloud.j2.yaml.

Depends-On: If074f87494a46305c990a0ea332c7b576d3c6ed8
Depends-On: Iab8aca2f1fcaba0c8f109717a4b3068f629c9aab
Partially-implements: blueprint composable-networks
Closes-bug: 1667104
Change-Id: I71339a6ac41133e95dbc3f93abb7a9fdeb0f2da0

7 years agoMerge "Copy scheduler configuration from service/ironic to services-docker/ironic"
Jenkins [Fri, 4 Aug 2017 13:20:11 +0000 (13:20 +0000)]
Merge "Copy scheduler configuration from service/ironic to services-docker/ironic"

7 years agoMerge "Fix up multipath docker indentation"
Jenkins [Fri, 4 Aug 2017 12:31:57 +0000 (12:31 +0000)]
Merge "Fix up multipath docker indentation"

7 years agoChange the directory for haproxy certs/keys to be service-specific
Juan Antonio Osorio Robles [Wed, 2 Aug 2017 07:27:15 +0000 (10:27 +0300)]
Change the directory for haproxy certs/keys to be service-specific

This moves the directories containing the certs/keys for haproxy one step
further inside the hierarchy. This way we will be able to bind-mount
this certificate into the container without bind-mounting any other
certs/keys from other services.

bp tls-via-certmonger-containers

Depends-On: Iba3adb9464a755e67c6f87d1233b3affa8be565a
Change-Id: I73df8d442b361cb5ef4e343b4ea2a198a5b95da9

7 years agoMerge "Adds environment file for ODL + SRIOV"
Jenkins [Fri, 4 Aug 2017 10:16:42 +0000 (10:16 +0000)]
Merge "Adds environment file for ODL + SRIOV"