apex-puppet-tripleo.git
7 years agoPuppet module to deploy MySQL bundle for HA
Damien Ciabrini [Mon, 10 Apr 2017 14:27:35 +0000 (14:27 +0000)]
Puppet module to deploy MySQL bundle for HA

This module is used by tripleo-heat-templates to configure and deploy
Kolla-based mysql containers managed by pacemaker.

We use short-lived containers that call pcs via puppet to create
the needed pacemaker resources, properties and constraints.

Co-Authored-By: Michele Baldesari <michele@acksyn.org>
Partial-Bug: #1692842

Depends-On: I44fbd7f89ab22b72e8d3fc0a0e3fe54a9418a60f
Depends-On: Ie9b7e7d2a3cec4b121915a17c1e809e4ec950e7f

Change-Id: I3b4d8ad2eec70080419882d5d822f78ebd3721ae

7 years agoMerge "Add support for autofencing to Pacemaker Remote."
Jenkins [Wed, 24 May 2017 18:21:45 +0000 (18:21 +0000)]
Merge "Add support for autofencing to Pacemaker Remote."

7 years agoMerge "Enable mistral to run under mod_wsgi"
Jenkins [Wed, 24 May 2017 18:21:28 +0000 (18:21 +0000)]
Merge "Enable mistral to run under mod_wsgi"

7 years agoEnable novajoin user on keystone profile
Juan Antonio Osorio Robles [Tue, 23 May 2017 06:24:36 +0000 (09:24 +0300)]
Enable novajoin user on keystone profile

If novajoin is enabled, the keystone profile should create its user.

bp tls-via-certmonger-containers

Change-Id: Ifb43b72cbf0180cf12e6d3584c92ae01ce5294e5

7 years agoMerge "TLS everywhere: Add resources for mongodb's TLS configuration"
Jenkins [Mon, 22 May 2017 07:52:07 +0000 (07:52 +0000)]
Merge "TLS everywhere: Add resources for mongodb's TLS configuration"

7 years agoMerge "Bad example in firewall.pp"
Jenkins [Fri, 19 May 2017 17:31:22 +0000 (17:31 +0000)]
Merge "Bad example in firewall.pp"

7 years agoMerge "Switch to overlay2 driver for storage"
Jenkins [Fri, 19 May 2017 17:30:47 +0000 (17:30 +0000)]
Merge "Switch to overlay2 driver for storage"

7 years agoBad example in firewall.pp
Cyril Lopez [Fri, 19 May 2017 10:16:13 +0000 (12:16 +0200)]
Bad example in firewall.pp

It was an error in the example to set firewall rule

Tested on newton.

Closes-Bug: 1691990

Change-Id: I091ad0305ac9d9fbb63853e46169fcaa6092456b
Signed-off-by: Cyril Lopez <cylopez@redhat.com>
7 years agoSwitch to overlay2 driver for storage
Dan Prince [Thu, 30 Mar 2017 17:24:55 +0000 (13:24 -0400)]
Switch to overlay2 driver for storage

This patch switches the default to the overlay2 storage driver and see
if it helps performance.

Background:

The loopback driver is not recommended for production. Most
other docker storage backends require extra disks (or partitions)
which we don't have on the root disk. Overlay seems to make the
most since for TripleO upgrades where we intend to update
in-place installations to use docker.

Co-Authored-By: Martin André <m.andre@redhat.com>
Change-Id: I6896a9b3e9dc3e269bf5b0dc753bf8c985482daf

7 years agoMerge "Update tox configuration"
Jenkins [Thu, 18 May 2017 17:16:44 +0000 (17:16 +0000)]
Merge "Update tox configuration"

7 years agoMerge "Update gitignore not to exclude fixture hieradata"
Jenkins [Thu, 18 May 2017 05:42:14 +0000 (05:42 +0000)]
Merge "Update gitignore not to exclude fixture hieradata"

7 years agoMerge "Handle duplicate/invalid entries in migration SSH inbound addresses"
Jenkins [Thu, 18 May 2017 00:13:41 +0000 (00:13 +0000)]
Merge "Handle duplicate/invalid entries in migration SSH inbound addresses"

7 years agoMerge "Disable SSH login for nova_migration user when migration over ssh is disabled."
Jenkins [Thu, 18 May 2017 00:13:19 +0000 (00:13 +0000)]
Merge "Disable SSH login for nova_migration user when migration over ssh is disabled."

7 years agoUpdate gitignore not to exclude fixture hieradata
Alex Schultz [Wed, 17 May 2017 20:45:35 +0000 (14:45 -0600)]
Update gitignore not to exclude fixture hieradata

The existing .gitignore is causing the hieradata we use for tests to be
excluded in git and our release tarballs. Lets adjust the gitignore not
to exclude the hiera files in spec/fixtures

Change-Id: Ic31687d0eb1c2e8acc92796d4c0eba096db8e533
Closes-Bug: #1691559

7 years agoUpdate tox configuration
Alex Schultz [Wed, 17 May 2017 16:59:30 +0000 (10:59 -0600)]
Update tox configuration

Update the tox configuration to pull in the openstack
upper-constraints.txt when running releasenotes. This will
fix the releasenotes job that is currently failing due to
a new version of sphinx. Additionally this change includes
updates from puppet-modulesync-configs.

Change-Id: Ie587bfde2367dfec796f1b07c01bba15d839a3b1
Related-Bug: #1691511

7 years agoTLS everywhere: Add resources for mongodb's TLS configuration
Juan Antonio Osorio Robles [Tue, 2 May 2017 13:15:02 +0000 (16:15 +0300)]
TLS everywhere: Add resources for mongodb's TLS configuration

bp tls-via-certmonger

Change-Id: I85dda29bcad686372a74bd7f094bfd62777a3032

7 years agoMerge "Use verify_on_create when creating pacemaker remote resources"
Jenkins [Tue, 16 May 2017 15:03:12 +0000 (15:03 +0000)]
Merge "Use verify_on_create when creating pacemaker remote resources"

7 years agoMerge "vhostuser socket dir shall be created for vhostuserclient mode"
Jenkins [Mon, 15 May 2017 17:46:17 +0000 (17:46 +0000)]
Merge "vhostuser socket dir shall be created for vhostuserclient mode"

7 years agovhostuser socket dir shall be created for vhostuserclient mode
Karthik S [Fri, 24 Mar 2017 09:33:41 +0000 (05:33 -0400)]
vhostuser socket dir shall be created for vhostuserclient mode

In order to support vhostuser client mode, a vhostuser_socket_dir
needs to be created with qemu:qemu g+w permissions.

Closes-Bug: #1675690

Co-Authored-By: Sanjay Upadhyay <supadhya@redhat.com>
Change-Id: I255f98c40869e7508ed01a03a96294284ecdc6a8
Signed-off-by: Karthik S <ksundara@redhat.com>
7 years agoMerge "Add support for Cinder "NAS secure" driver params"
Jenkins [Thu, 11 May 2017 16:49:18 +0000 (16:49 +0000)]
Merge "Add support for Cinder "NAS secure" driver params"

8 years agoMerge "Migrates OpenDaylight to official repo"
Jenkins [Mon, 8 May 2017 20:26:04 +0000 (20:26 +0000)]
Merge "Migrates OpenDaylight to official repo"

8 years agoUse verify_on_create when creating pacemaker remote resources
Michele Baldessari [Sat, 6 May 2017 15:40:24 +0000 (17:40 +0200)]
Use verify_on_create when creating pacemaker remote resources

We currently create remote resources without waiting for their creation.
This leads to the following potential race (spotted by Marian Mkrcmari):
- On Step1 pacemaker bootstrap node creates the resource but the remote
  resource is not yet created
- Step1 completes and Step2 starts
- On Step2 the remote node sets a property (or calls pcs cib) but the
  remote is not yet set up so 'pcs cluster cib' will fail there with:

    (err): Could not evaluate: backup_cib: Running: /usr/sbin/pcs cluster
    cib /var/lib/pacemaker/cib/puppet-cib-backup20170506-15994-1swnk1i failed
    with code: 1 ->

Note that when verify_on_create is set to true we are not using the cib
dump/push mechanism. That is fine because we create the remotes on
step1 and the dump/push mechanism is only needed starting from step2
when multiple nodes set cluster properties at the same time.

Tested by Marian Mkrcmari successfully as well.

Closes-Bug: #1689028

Change-Id: I764526b3f3c06591d477cc92779d83a19802368e
Depends-On: I1db31dcc92b8695ab0522bba91df729b37f34e0f

8 years agoEnable mistral to run under mod_wsgi
Brad P. Crochet [Thu, 20 Apr 2017 13:33:29 +0000 (09:33 -0400)]
Enable mistral to run under mod_wsgi

Mistral should run under mod_wsgi. Enable that.

Change-Id: I99f83c35eaa892c10deb63e199d22a43f06f5dcc
Depends-On: I61199f53d7e32fcb3d068ccaf548a836b5bb58e9

8 years agoMigrates OpenDaylight to official repo
Tim Rozet [Fri, 5 May 2017 17:41:34 +0000 (13:41 -0400)]
Migrates OpenDaylight to official repo

ODL repo has been moved from a personal github account into official ODL
repo.

Change-Id: I1248c22acc09962ef788ca19ee87b6bda8ef8450
Signed-off-by: Tim Rozet <trozet@redhat.com>
8 years agoRemove limits for redis in /etc/security/limits.d
Michele Baldessari [Fri, 5 May 2017 10:29:39 +0000 (12:29 +0200)]
Remove limits for redis in /etc/security/limits.d

Now that puppet-redis supports ulimit for cluster managed redis (via
https://github.com/arioch/puppet-redis/pull/192), we need to remove the
file snippet as otherwise we will get a duplicate resource error.

We will need to create a THT change that at the very least sets the
redis::managed_by_cluster_manager key to true so that
/etc/security/limits.d/redis.conf gets created.
We also add code to not break backwards compatibility with the old hiera
key.

Change-Id: I4ffccfe3e3ba862d445476c14c8f2cb267fa108d
Partial-Bug: #1688464

8 years agoHandle duplicate/invalid entries in migration SSH inbound addresses
Oliver Walsh [Fri, 5 May 2017 00:30:21 +0000 (01:30 +0100)]
Handle duplicate/invalid entries in migration SSH inbound addresses

An error (e.g a typo) in a custom tripleo-heat-templates environment
file could lead to an invalid match block in /etc/ssh/sshd_config.
SSH fails-safe and refuses all logins in this case.

This change validates the migration_ssh_localaddrs parameter is an
array of IP addresses and removes and duplicate entries.

Change-Id: Ibcf144d960fe52f0eab0d5015bd30cf7c1e37e25
Closes-Bug: #1688308

8 years agoDisable SSH login for nova_migration user when migration over ssh is disabled.
Oliver Walsh [Thu, 4 May 2017 19:21:51 +0000 (20:21 +0100)]
Disable SSH login for nova_migration user when migration over ssh is disabled.

If migration over ssh is enabled, and then later disabled, the ssh config
for the nova_migration user remains intact. This change clobbers the migration
SSH key to disable login when it is not necessary.

Change-Id: Icc6d5d4f4671b3525a731d334ca6fa7c5419dac3
Closes-Bug: #1688321

8 years agoMerge "Restrict nova migration ssh tunnel"
Jenkins [Thu, 4 May 2017 17:30:30 +0000 (17:30 +0000)]
Merge "Restrict nova migration ssh tunnel"

8 years agoAdd support for Cinder "NAS secure" driver params
Alan Bishop [Thu, 4 May 2017 16:27:27 +0000 (12:27 -0400)]
Add support for Cinder "NAS secure" driver params

Add ability to set Cinder's nas_secure_file_operations and
nas_secure_file_permissions driver parameters. Two sets of identically
named parameters are implemented by Cinder's NFS and NetApp back end
drivers.

The ability to control these parameters is crucial for supporting deployments
that require non-default values.

Partial-Bug: #1688332
Depends-On: Id92cfd4190de8687d4731cf301f2df0bde1ba7d9
Change-Id: I76e2ce10acf7b671be6a2785829ebb3012b79308

8 years agoMerge "MySQL client: Make CA file configurable"
Jenkins [Thu, 4 May 2017 10:50:09 +0000 (10:50 +0000)]
Merge "MySQL client: Make CA file configurable"

8 years agoMerge "snmp: remove useless parameter for binding"
Jenkins [Wed, 3 May 2017 21:14:50 +0000 (21:14 +0000)]
Merge "snmp: remove useless parameter for binding"

8 years agoRestrict nova migration ssh tunnel
Oliver Walsh [Wed, 19 Apr 2017 13:39:42 +0000 (14:39 +0100)]
Restrict nova migration ssh tunnel

This change enhances the security of the migration ssh tunnel:
- The ssh authorized_keys file is only writeable by root.
- Creates a new user for migration instead of using root/nova.
- Disables SSH forwarding for this user.
- Optionally restricts the networks that this user can connect from.
- Uses an ssh wrapper command to whitelist the commands that this user can run
  over ssh.

Requires the openstack-nova-migration package from
https://review.rdoproject.org/r/6327

bp tripleo-cold-migration

Change-Id: Idb56acd1e1ecb5a5fd4d942969be428cc9cbe293

8 years agoMerge "IPv6 VIP addresses need to be /128"
Jenkins [Wed, 3 May 2017 19:43:11 +0000 (19:43 +0000)]
Merge "IPv6 VIP addresses need to be /128"

8 years agoMySQL client: Make CA file configurable
Juan Antonio Osorio Robles [Wed, 3 May 2017 09:54:38 +0000 (12:54 +0300)]
MySQL client: Make CA file configurable

It used to be hardcoded to use the OpenSSL default CA Bundle, however,
this will be changed in t-h-t.

Change-Id: I75bdaf71d88d169e64687a180cb13c1f63418a0f

8 years agoIPv6 VIP addresses need to be /128
Michele Baldessari [Wed, 26 Apr 2017 09:40:51 +0000 (11:40 +0200)]
IPv6 VIP addresses need to be /128

We currently hardcode /64 as our VIP addresses when using IPv6.
The problem with this is that some server code might bind to that
IP as a source address when doing inter-cluster communication
(rabbitmq/galera for example). So when the VIP moves there will
be effectively a network outage between the nodes, which should not
happen.

Likely this was hardcoded to /64 because the RA IPaddr2 needs a nic
parameter when /128 is specified. This is due to:
https://bugzilla.redhat.com/show_bug.cgi?id=1445628

We also make sure we use the ipv6_addrlabel option set to 99 so that
they will never be used as source ip addresses.

Depends-On: I7fcf15a00aedbdcfb21db501ad46c69fb97ec30c
Partial-Bug: #1686357

Change-Id: Ibefde870512ad1e03ff12f7aea91b3734f03f96f
Co-Authored-By: Sofer Athlan-Guyot <sathlang@redhat.com>
Co-Authored-By: Marios Andreou <mandreou@redhat.com>
Co-Authored-By: Damien Ciabrini <dciabrin@redhat.com>
8 years agoMerge "firewall: generally accept "jump" param and use tripleo:firewall for log rule"
Jenkins [Wed, 3 May 2017 07:39:07 +0000 (07:39 +0000)]
Merge "firewall: generally accept "jump" param and use tripleo:firewall for log rule"

8 years agosnmp: remove useless parameter for binding
Emilien Macchi [Tue, 2 May 2017 13:53:51 +0000 (09:53 -0400)]
snmp: remove useless parameter for binding

Binding is now done in THT via Hiera directly, so users can change the
option more easily.

Depends-On: Iccf0a8d35cc05d34272c078c97a5dddfb8e7d614
Change-Id: I9d5fd152bb73ea54c4d0d3bab862f11eaa4ebd79
Closes-Bug: #1687628

8 years agoMerge "Fix wrong notify in swift proxy profile"
Jenkins [Thu, 27 Apr 2017 17:33:56 +0000 (17:33 +0000)]
Merge "Fix wrong notify in swift proxy profile"

8 years agoMerge "Add linuxbridge agent profile"
Jenkins [Thu, 27 Apr 2017 15:59:58 +0000 (15:59 +0000)]
Merge "Add linuxbridge agent profile"

8 years agoMerge "Include base apache module in tls_proxy resource"
Jenkins [Thu, 27 Apr 2017 14:52:12 +0000 (14:52 +0000)]
Merge "Include base apache module in tls_proxy resource"

8 years agoFix wrong notify in swift proxy profile
Juan Antonio Osorio Robles [Thu, 27 Apr 2017 07:23:36 +0000 (10:23 +0300)]
Fix wrong notify in swift proxy profile

the TLS proxy was notifying neutron::server instead of swift proxy.

Change-Id: I212978c107a75209d5b7c266e608eb9a9e9cdc76

8 years agoInclude base apache module in tls_proxy resource
Juan Antonio Osorio Robles [Thu, 27 Apr 2017 06:56:13 +0000 (09:56 +0300)]
Include base apache module in tls_proxy resource

Other services include it by using the vhost resource from openstacklib.
If we include a service (such as swift-proxy) that uses the tls_proxy
resource, and we do so in a separate node or in its own container, it
will fail since the base apache module hadn't been included.

Change-Id: I0167e08b0b652618d8a1af792376bcf02c8fcd82

8 years agoAdd support for autofencing to Pacemaker Remote.
Chris Jones [Tue, 25 Apr 2017 15:05:27 +0000 (16:05 +0100)]
Add support for autofencing to Pacemaker Remote.

We now configure stonith devices for Pacemaker Remote nodes.

Change-Id: I87c60bd56feac6dedc00a3c458b805aa9b71d9ce
Depends-On: Ifb4d19a6b9920b0e340555d6441878c7234eb197
Partial-Bug: #1686115

8 years agoAdd a flag to rabbitmq so that we can deploy with ha-mode: all again
Michele Baldessari [Wed, 26 Apr 2017 08:12:50 +0000 (10:12 +0200)]
Add a flag to rabbitmq so that we can deploy with ha-mode: all again

In change Ib62001c03e1e08f58cf0c6e0ba07a8879a584084 we switched the
rabbitmq queues HA mode from ha-all to ha-exactly. While this gives us a
nice performance boost with rabbitmq, it makes rabbit less resilient to
network glitches as we painfully found out via
https://bugzilla.redhat.com/show_bug.cgi?id=1441635.

Will propose another THT change to actually change the default to
-1 so we get this ha-mode:all by default.

Change-Id: I9a90e71094b8d8d58b5be0a45a2979701b0ac21c
Partial-Bug: #1686337
Co-Authored-By: Damien Ciabrini <dciabrin@redhat.com>
Co-Authored-By: John Eckersberg <jeckersb@redhat.com>
8 years agoMerge "Enable internal network TLS for etcd"
Jenkins [Tue, 25 Apr 2017 17:33:53 +0000 (17:33 +0000)]
Merge "Enable internal network TLS for etcd"

8 years agoMerge "Update puppet-etcd version"
Jenkins [Tue, 25 Apr 2017 13:35:02 +0000 (13:35 +0000)]
Merge "Update puppet-etcd version"

8 years agoMerge "Add support for Redfish hardware in Ironic"
Jenkins [Tue, 25 Apr 2017 11:29:16 +0000 (11:29 +0000)]
Merge "Add support for Redfish hardware in Ironic"

8 years agoMerge "Include zaqar apache module"
Jenkins [Tue, 25 Apr 2017 09:53:09 +0000 (09:53 +0000)]
Merge "Include zaqar apache module"

8 years agoMerge "Dell SC: Add secondary DSM support"
Jenkins [Tue, 25 Apr 2017 01:58:09 +0000 (01:58 +0000)]
Merge "Dell SC: Add secondary DSM support"

8 years agoUpdate puppet-etcd version
Bogdan Dobrelya [Mon, 24 Apr 2017 16:07:00 +0000 (18:07 +0200)]
Update puppet-etcd version

Include the yaml config file for containers

Change-Id: I3ad463217ed3f2d6374627248236b274cfed72fb
Signed-off-by: Bogdan Dobrelya <bdobreli@redhat.com>
8 years agoAdd support for Redfish hardware in Ironic
Dmitry Tantsur [Mon, 24 Apr 2017 12:48:28 +0000 (14:48 +0200)]
Add support for Redfish hardware in Ironic

Part of blueprint redfish-support

Depends-On: Icd065cec7114fc026b658ede0d78be2e777c15aa
Change-Id: Ib14f87800ae7657cf6176a4820248a2ce048241d

8 years agoMerge "Enable setting SubjectaltNames for haproxy and httpd certs"
Jenkins [Mon, 24 Apr 2017 12:17:17 +0000 (12:17 +0000)]
Merge "Enable setting SubjectaltNames for haproxy and httpd certs"

8 years agoMove ceilometer upgrade re-run out of collector
Pradeep Kilambi [Tue, 18 Apr 2017 23:14:16 +0000 (19:14 -0400)]
Move ceilometer upgrade re-run out of collector

Since collector is deprecated, lets move this out of collector.pp
so it gets run and resource types are created appropriately even
when collector is not included.

Closes-bug: #1676961

Change-Id: I32445a891c34f519ab16dcecc81993f8909f6481

8 years agoMerge "Allow to configure haproxy daemon's status"
Jenkins [Fri, 21 Apr 2017 13:39:13 +0000 (13:39 +0000)]
Merge "Allow to configure haproxy daemon's status"

8 years agoMerge "Cover gnocchi api step 4 and 5"
Jenkins [Fri, 21 Apr 2017 13:10:25 +0000 (13:10 +0000)]
Merge "Cover gnocchi api step 4 and 5"

8 years agoMerge "Add resource profile for vmware nsx_v3"
Jenkins [Fri, 21 Apr 2017 13:01:29 +0000 (13:01 +0000)]
Merge "Add resource profile for vmware nsx_v3"

8 years agoMerge "Add ML2 configuration for Bagpipe BGPVPN extension"
Jenkins [Fri, 21 Apr 2017 12:55:50 +0000 (12:55 +0000)]
Merge "Add ML2 configuration for Bagpipe BGPVPN extension"

8 years agoMerge "Refactor SSHD config to allow both SSHD options and banner/motd to be set"
Jenkins [Fri, 21 Apr 2017 05:00:27 +0000 (05:00 +0000)]
Merge "Refactor SSHD config to allow both SSHD options and banner/motd to be set"

8 years agoMerge "Update UI language list"
Jenkins [Fri, 21 Apr 2017 00:06:41 +0000 (00:06 +0000)]
Merge "Update UI language list"

8 years agoInclude zaqar apache module
Thomas Herve [Tue, 21 Mar 2017 08:51:26 +0000 (09:51 +0100)]
Include zaqar apache module

This includes the Zaqar apache module, allowing to run Zaqar behind
httpd.

Depends-On: I69b923dd76a60e9ec786cae886c137ba572ec906
Change-Id: Ib52144e5877d9293057713d6bdca557724baad5c

8 years agoMerge "Haproxy: When using TLS everywhere, use verifyhost for the balancermembers"
Jenkins [Thu, 20 Apr 2017 09:26:53 +0000 (09:26 +0000)]
Merge "Haproxy: When using TLS everywhere, use verifyhost for the balancermembers"

8 years agoRefactor SSHD config to allow both SSHD options and banner/motd to be set
Oliver Walsh [Tue, 18 Apr 2017 11:51:36 +0000 (12:51 +0100)]
Refactor SSHD config to allow both SSHD options and banner/motd to be set

In https://review.openstack.org/#/c/444622/7 the sshd_options and banner/motd
are mutually exclusive. This patch, and the next patchset of that review,
resolves the conflict.

Related-Bug: 1668543

Change-Id: I1d09530d69e42c0c36311789166554a889e46556

8 years agoCover gnocchi api step 4 and 5
Alex Schultz [Wed, 12 Apr 2017 19:53:44 +0000 (13:53 -0600)]
Cover gnocchi api step 4 and 5

Update the gnocchi api to expose the redis information as a class
parameter so it can be tested correctly.

Change-Id: I075b4af5e7bb35f90f7b82f8fb1b6d6ad6363b71

8 years agoEnsure /etc/docker/daemon.json
Dan Prince [Wed, 19 Apr 2017 21:55:54 +0000 (16:55 -0500)]
Ensure /etc/docker/daemon.json

A recent Centos docker packaging change removed the default
/etc/docker/daemon.json file. As such we need to create an empty
json file if none exists before running Augeas to configure
the settings.

Change-Id: Ibfe04b468639002f55da7bb65d2606f730c700b7
Closes-bug: #1684297

8 years agoDell SC: Add secondary DSM support
rajinir [Mon, 10 Apr 2017 16:46:03 +0000 (11:46 -0500)]
Dell SC: Add secondary DSM support

Adds support for a secondary DSM in case the primary becomes
unavailable.

Change-Id: Ibf8c333f62556d421d67c853f1f0740d7f9985bf
Depends-On: I331466e4f254b2b8ff7891b796e78cd30c2c87f7

8 years agoAllow to configure haproxy daemon's status
Michele Baldessari [Wed, 19 Apr 2017 10:56:46 +0000 (12:56 +0200)]
Allow to configure haproxy daemon's status

Currently we hard-code the fact that haproxy starts as a daemon.
When running haproxy in a container we need this to be configurable
because the haproxy process will be pid number 1.

We are not changing the current semantics which have the 'daemon'
option always set, but we are allowing its disabling.

Change-Id: I51c482b70731f15fee4025bbce14e46a49a49938

8 years agoMerge "Ensure we configure ssl.conf"
Jenkins [Wed, 19 Apr 2017 10:45:55 +0000 (10:45 +0000)]
Merge "Ensure we configure ssl.conf"

8 years agoAdd linuxbridge agent profile
Bartosz Stopa [Wed, 19 Apr 2017 07:18:27 +0000 (09:18 +0200)]
Add linuxbridge agent profile

Add a tripleo profile for neutron linuxbridge agent configuration.

Change-Id: Ie3ac03052f341c26735b423701e1decf7233d935
Partial-Bug: #1652211

8 years agoMerge "Create bigswitch agent profile"
Jenkins [Wed, 19 Apr 2017 03:11:07 +0000 (03:11 +0000)]
Merge "Create bigswitch agent profile"

8 years agoMerge "Added release note for "Support for external swift proxy""
Jenkins [Tue, 18 Apr 2017 19:14:35 +0000 (19:14 +0000)]
Merge "Added release note for "Support for external swift proxy""

8 years agoEnsure we configure ssl.conf
Lukas Bezdicka [Thu, 13 Apr 2017 17:21:45 +0000 (19:21 +0200)]
Ensure we configure ssl.conf

Every time we call apache module regardless of using SSL we have to
configure mod_ssl from puppet-apache or we'll hit issue during package
update. File /etc/httpd/conf.d/ssl.conf from mod_ssl package contains
Listen 443 while apache::mod::ssl just configures SSL bits but does not
add Listen. If the apache::mod::ssl is not included the ssl.conf file is
removed and recreated during mod_ssl package update. This causes
conflict on port 443.

Change-Id: Ic5a0719f67d3795a9edca25284d1cf6f088073e8
Related-Bug: 1682448
Resolves: rhbz#1441977

8 years agoEnable setting SubjectaltNames for haproxy and httpd certs
Juan Antonio Osorio Robles [Tue, 18 Apr 2017 12:48:02 +0000 (15:48 +0300)]
Enable setting SubjectaltNames for haproxy and httpd certs

This enables setting the subjectAltNames for HAProxy and httpd certs.
These will eventually replace the usage of many certs, to have instead
just one that has several subjectAltNames.

Change-Id: Icd152c8e0389b6a104381ba6ab4e0944e9828ba3

8 years agoAdded release note for "Support for external swift proxy"
Luca Lorenzetto [Tue, 18 Apr 2017 12:03:54 +0000 (14:03 +0200)]
Added release note for "Support for external swift proxy"

Change-Id: I7feac65bf814099ab591b473be962e64dec85cbd

8 years agoHaproxy: When using TLS everywhere, use verifyhost for the balancermembers
Juan Antonio Osorio Robles [Tue, 18 Apr 2017 11:49:09 +0000 (14:49 +0300)]
Haproxy: When using TLS everywhere, use verifyhost for the balancermembers

This checks that the subjectAltName in the backend server's certificate
matches the server's name that was intended to be used.

Change-Id: If1c61e1becf9cc84c9b18835aef1eaaa8c0d4341

8 years agoMerge "Allow setting of keepalived router ID"
Jenkins [Tue, 18 Apr 2017 10:59:03 +0000 (10:59 +0000)]
Merge "Allow setting of keepalived router ID"

8 years agoMerge "HAproxy/heat_api: increase timeout to 10m"
Jenkins [Tue, 18 Apr 2017 06:58:02 +0000 (06:58 +0000)]
Merge "HAproxy/heat_api: increase timeout to 10m"

8 years agoHAproxy/heat_api: increase timeout to 10m
Emilien Macchi [Mon, 17 Apr 2017 15:30:23 +0000 (11:30 -0400)]
HAproxy/heat_api: increase timeout to 10m

Default timeout is 2min but it doesn't reflect the rpc_response_timeout
value that we set in THT and instack-undercloud, which is 600 (10 min).

In some cases (in low-memory environments), Heat needs more than 2
minutes to reply to the client, when deploying the overcloud.
It makes sense to increase the timeout to the value of rpc_timeout to
give a chance to Heat to reply to the client, otherwise HAproxy will
kill the connection and send 504 to the client.

Depends-On: I9669d40d86d762101734704fcef153e360767690
Change-Id: I32c71fe7930c8798d306046d6933e4b20c22740c
Related-Bug: 1666072

8 years agoMerge "Support for external swift proxy"
Jenkins [Mon, 17 Apr 2017 18:06:40 +0000 (18:06 +0000)]
Merge "Support for external swift proxy"

8 years agoMerge "Move ceilometer wsgi to step 3"
Jenkins [Sat, 15 Apr 2017 02:57:28 +0000 (02:57 +0000)]
Merge "Move ceilometer wsgi to step 3"

8 years agoMerge "Move gnocchi wsgi configuration to step 3"
Jenkins [Sat, 15 Apr 2017 02:54:56 +0000 (02:54 +0000)]
Merge "Move gnocchi wsgi configuration to step 3"

8 years agoMerge "Dell SC: Add exclude_domain_ip option"
Jenkins [Fri, 14 Apr 2017 23:59:15 +0000 (23:59 +0000)]
Merge "Dell SC: Add exclude_domain_ip option"

8 years agoSupport for external swift proxy
Luca Lorenzetto [Fri, 14 Apr 2017 08:45:57 +0000 (10:45 +0200)]
Support for external swift proxy

Users may have an external swift proxy already available (i.e. radosgw
from already existing ceph, or hardware appliance implementing swift
proxy). With this change user may specify an environment file that
registers the specified urls as endpoint for the object-store service.
The internal swift proxy is left as unconfigured.

Change-Id: Ia568c3a5723d8bd8c2c37dbba094fc8a83b9d67e

8 years agoMerge "Make install of kolla optional on the undercloud"
Jenkins [Thu, 13 Apr 2017 22:15:15 +0000 (22:15 +0000)]
Merge "Make install of kolla optional on the undercloud"

8 years agoMerge "etcd: Make HAProxy terminate TLS connections"
Jenkins [Thu, 13 Apr 2017 15:20:28 +0000 (15:20 +0000)]
Merge "etcd: Make HAProxy terminate TLS connections"

8 years agoAllow setting of keepalived router ID
Thomas Herve [Wed, 12 Apr 2017 16:02:46 +0000 (18:02 +0200)]
Allow setting of keepalived router ID

By default the undercloud and the overcloud share virtual_router_id
definition, leading to errors like "ip address associated with VRID not
present in received packet". This allows setting the range for the IDs.

Change-Id: I0c822777824b469b0f8ef0f31b3708fe47d5b2d7

8 years agoDell SC: Add exclude_domain_ip option
rajinir [Mon, 10 Apr 2017 16:41:29 +0000 (11:41 -0500)]
Dell SC: Add exclude_domain_ip option

This option allows users to exclude some fault domains.
Otherwise all domains are returned.

Change-Id: I6eb2bcc7db003a5eebd3924e3e4eb44e35f60483
Depends-On: I8ac91e6720e52da9cf7480f80bcfb456bf0c2433

8 years agoMake install of kolla optional on the undercloud
Martin André [Wed, 12 Apr 2017 16:06:15 +0000 (18:06 +0200)]
Make install of kolla optional on the undercloud

This defaults to 'True' to keep backward compatibility and can be
disabled by setting 'enable_container_images_built' to false in
undercloud.conf.

Depends-On: Ia3379cf66b1d6b180def69c2a5b22b2602baacef
Change-Id: I33e7e9a6a3865fed38f7ed6490455457da67782b

8 years agoMove gnocchi wsgi configuration to step 3
Alex Schultz [Wed, 12 Apr 2017 16:34:07 +0000 (10:34 -0600)]
Move gnocchi wsgi configuration to step 3

We configure apache in step3 so we need to configure the gnocchi api in
step 3 as well to prevent unnecessary service restarts during updates.

Change-Id: I30010c9cf0b0c23fde5d00b67472979d519a15be
Related-Bug: #1664418

8 years agoMove ceilometer wsgi to step 3
Alex Schultz [Wed, 12 Apr 2017 16:03:22 +0000 (10:03 -0600)]
Move ceilometer wsgi to step 3

Apache is configured in step 3 so if we configure ceilometer in step 4,
the configuration is removed on updates. We need to configure it in step
3 with the other apache services to ensure we don't have issues on
updates.

Change-Id: Icc9d03cd8904c93cb6e17f662f141c6e4c0bf423
Related-Bug: #1664418

8 years agoMerge "Stop SSHD profile clobbering SSH client config"
Jenkins [Wed, 12 Apr 2017 15:58:21 +0000 (15:58 +0000)]
Merge "Stop SSHD profile clobbering SSH client config"

8 years agoAdd ML2 configuration for Bagpipe BGPVPN extension
Ricardo Noriega [Wed, 12 Apr 2017 10:26:19 +0000 (12:26 +0200)]
Add ML2 configuration for Bagpipe BGPVPN extension

Change-Id: I9e1a56782e258fb6982b70d9a07f35808f2b2de5
Depends-On: Ic975ec1d6b2bf6e6bd28b47ba9dd2a3ae629d149
Signed-off-by: Ricardo Noriega <rnoriega@redhat.com>
8 years agoMerge "Ensure directory exists for certificates for httpd"
Jenkins [Wed, 12 Apr 2017 10:14:09 +0000 (10:14 +0000)]
Merge "Ensure directory exists for certificates for httpd"

8 years agoEnable internal network TLS for etcd
Feng Pan [Fri, 7 Apr 2017 20:24:10 +0000 (16:24 -0400)]
Enable internal network TLS for etcd

bp secure-etcd

Change-Id: I0759deef7cbcf13b9056350e92f01afd33e9c649
Signed-off-by: Feng Pan <fpan@redhat.com>
8 years agoStop SSHD profile clobbering SSH client config
Oliver Walsh [Tue, 11 Apr 2017 20:42:44 +0000 (21:42 +0100)]
Stop SSHD profile clobbering SSH client config

Including the ::ssh manifest will manage both client and server config.
Managing the client config was not intended and will clobber the OS
default config with the puppet ssh moduled defaults.

Follow up for https://review.openstack.org/443113 where I found the issue after
the changes merged.

Change-Id: I6329f5ebbe8fc3950449e325e56293872d11e1b5
Related-Bug: 1668543

8 years agoEnsure directory exists for certificates for httpd
Juan Antonio Osorio Robles [Fri, 24 Mar 2017 09:31:12 +0000 (11:31 +0200)]
Ensure directory exists for certificates for httpd

We used to rely on a standard directory for the certificates and keys
that are requested by certmonger. However, given the approach we plan to
take for containers that's described in the blueprint, we need to use
service-specific directories for the certs/keys, since we plan to
bind-mount these into the containers, and we don't want to bind mount
any keys/certs from other services.

Thus, we start by creating this directories if they don't exist in the
filesystem and adding the proper selinux labels.

bp tls-via-certmonger-containers

Change-Id: I0b71902358b754fa8bd7fdbb213479503c87aa46

8 years agoUpdate UI language list
Julie Pichon [Tue, 11 Apr 2017 10:41:49 +0000 (11:41 +0100)]
Update UI language list

Change-Id: I848b3cc747f1be06aeda57ba15d4ec557c23ad46
Depends-On: Idf3d82058d87d9c8a3b6d8973d5166043dad2252

8 years agoMerge "Add registry_mirror to base::docker profile"
Jenkins [Tue, 11 Apr 2017 06:10:03 +0000 (06:10 +0000)]
Merge "Add registry_mirror to base::docker profile"

8 years agoMerge "Use docker profile in docker_registry"
Jenkins [Tue, 11 Apr 2017 02:25:15 +0000 (02:25 +0000)]
Merge "Use docker profile in docker_registry"

8 years agoetcd: Make HAProxy terminate TLS connections
Juan Antonio Osorio Robles [Mon, 10 Apr 2017 13:09:51 +0000 (16:09 +0300)]
etcd: Make HAProxy terminate TLS connections

When TLS is enabled for the internal network, HAProxy needs to handle
etcd's TLS termination. Else it will use plain text.

bp secure-etcd

Change-Id: I20651240edcff0953741d4e8e01fa9a7ab185863

8 years agoMerge "Move etcd to step 2"
Jenkins [Mon, 10 Apr 2017 13:45:40 +0000 (13:45 +0000)]
Merge "Move etcd to step 2"