aboutsummaryrefslogtreecommitdiffstats
path: root/puppet/manifests
AgeCommit message (Collapse)AuthorFilesLines
2015-04-28Add RabbitMQ as a Pacemaker resource when EnablePacemakerGiulio Fidente1-28/+46
Change-Id: I43a74c1db324144d33e96a94cb718db30e0fd243
2015-04-27Configures haproxy as a Pacemaker resourceGiulio Fidente1-0/+3
Change-Id: I6bf5ada5a5298f4079594f3cc8b01ac0ef85876e
2015-04-27Switch VIP management from Keepalived to PacemakerGiulio Fidente1-0/+11
Change-Id: I45511569fda6b00ca35b1e590537a29271e56ce0 Depends-On: I98b9b3dbc48009ce255d964ac580e1a31f279f1e
2015-04-21Parameter to enable/disable controller swift storeDan Prince1-12/+14
Adds a new ControllerEnableSwiftStorage parameter that can be used to enable/disable use of the contoller node as a Swift storage node. Change-Id: Ic54144f4a46a671818c2f12e419cfa619b0dc1f9
2015-04-21Add option to enable ceph storage on controllerDan Prince1-0/+5
This patch adds a new ControllerEnableCephStorage option which can be used to install and configure Ceph storage (OSD) on the controller node. The default is to have this disabled by default (this is probably a more production like setting). The motivation for this change is to help facilitate CI jobs which actually use Ceph. Right now we have an issue where once the Heat stack finishes Ceph is configured and ready, but Cinder volume (required by our CI devtest_overcloud.sh test) may or may not have had enough time to recognize the amount of storage on the remote Ceph storage nodes. Waiting another periodic cycle for Cinder volume to recognize the actual amount of storage on the remote OSD nodes would work but there isn't a good way to do this ATM. The right solution here is probably to implement Heat breakpoints in our CI. As we haven't quite landed that change, another option is to simply make the controller node also be a Ceph storage node. Since this runs as "step 2" within the controller it ensures that the OSD will be available and thus Cinder volume will register the correct amount of storage on startup. Enabling this feature also matches what we do with Swift storage on the Controller (although we should provide an option to actually disable this as well). Change-Id: Ic47d028591edbaab83a52d7f38283d7805b63042
2015-04-21Merge "Perform basic setup of Pacemaker cluster using puppet-pacemaker"Jenkins1-15/+23
2015-04-21Merge "Add support for Redis configuration"Jenkins1-1/+21
2015-04-20Perform basic setup of Pacemaker cluster using puppet-pacemakerGiulio Fidente1-15/+23
Depends-On: Ia1bbf53c674e34ba7c70249895b106ec0af3c249 Change-Id: Ifa9f579d26a3cba9f8705226984c7b987ae0ad1c
2015-04-16Add support for Redis configurationYanis Guenane1-1/+21
Add support for Redis configuration on the overcloud controller role. Change-Id: I917ff1e7c0abf9d76b9939a97978e858268deac2 Depends-On: I80a6c284af9eceb6b669a03c5d93256261523331
2015-04-16puppet: add missing ::glance include on controllerEmilien Macchi1-0/+1
On Controller node, we also need to include ::glance if we want to have common Glance bits (packaging included). This is a Puppet best practice. Change-Id: I967c06b2c78d8f3aa5fa984b518d34c813426a2e
2015-04-16Merge "Do not downcase IP addresses"Jenkins1-2/+2
2015-04-16Merge "Perform basic setup of pacemaker cluster on controllers"Jenkins1-0/+17
2015-04-15puppet: Ceilometer use MongoDB datastore backendEmilien Macchi1-9/+5
This patch configure Ceilometer to use MongoDB backend. Change-Id: I22be0e22e7a3991ebd2d3aa7d14c518418a2458a
2015-04-15Fix MongoDB setup on nonha setupYanis Guenane1-6/+4
Currently a replset parameter is set in mongodb.conf no matter if we are in a ha or nonha setup. This install fine, but on a nonha setup it prevents any program from using MongoDB, since no replset has been initialized. It generates the following error when a program tries to use it : not master and slaveOk=false To prevent this issue a replicatset is initialized in both ha and nonha setup, this way if another MongoDB node is added to the pool, it will be able to attach automatically Change-Id: I65e3f1ad35cb0cd31f6771444a0cffdf7569222f
2015-04-14Do not downcase IP addressesGiulio Fidente1-2/+2
Change-Id: I0655b7cae2c436944833894bf9837877b3a69878
2015-04-14puppet: implement MongoDB on controller nodesEmilien Macchi1-0/+15
This patch aims to configure MongoDB server on controller nodes with Puppet. It also create a default replicaset for Ceilometer, so MongoDB can be highly available when multiple controllers are run. Change-Id: I3c1ff06ebc3c9dac44fc790caaea711d0eba4bb7
2015-04-14Perform basic setup of pacemaker cluster on controllersGiulio Fidente1-0/+17
Change-Id: Ia2e4eae619ca95c0f417f713676732eb4f01304b Depends-On: I9563eec0a2266deb2ebef2e3d76ae89d39b2be29
2015-04-10MariaDB bind on correct addressJiri Stransky1-0/+7
Despite passing bind-address for MariaDB in overcloud_controller.pp correctly, it was always trying to bind on 0.0.0.0. The problem is caused by Galera's config file (we install Galera into the image even though we don't use it yet). Galera's default config file contains override of the bind-address value to 0.0.0.0, and the setting from galera.cnf took precendence over what was in server.cnf. The mariadb-galera-server package assumes that the main config happens in galera.cnf and it ships an almost empty server.cnf. We now have an EnableGalera param, when it's set to true the mysql module will manage galera.cnf instead of server.cnf, overriding the default values from galera.cnf and fixing the issue. Change-Id: I7c2fd41d41dcf5eb4ee8b1dbd74d60cc2cabeed9 Closes-Bug: #1442256
2015-04-09Pass in libvirt_rbd_secret_key for nova computeJiri Stransky1-1/+5
Passing the key explicitly into nova::compute::rbd means that Puppet will not attempt to fetch the key using `ceph auth get-key <keyring>`, having these effects: * One reason for compute node to have access to the client.admin key is gone (in current implementation it does have access to the key, but this change is a step towards removing it). * Ceph cluster doesn't have to be running at the time when Puppet runs on compute node, meaning we don't have to serialize things more than we do now. Also adding the ComputeCephDeployment as a dependency of ComputePostDeployment, otherwise the hiera file it creates might be created *after* Puppet configuration happens on compute nodes, and the values it provides would be missing during the Puppet run on the compute nodes. Change-Id: Id3166e6d5f01d18ec8a5033398bb511f4321a5e8 Depends-On: I70da06159c0d3c6fa204b5f7a468909ffab4d633 Partial-Bug: #1439949
2015-04-03Remove enforced deps on RabbitMQ resources before configGiulio Fidente1-4/+0
This should have been removed with change I1bb8ee15d361638d77c5df7f8c03561c34f4c88f Change-Id: I20d4099aabe5ae9f89db45fd3db585067cab01f5
2015-04-02Merge "Restructure Ceph/Puppet params to reflect changes in puppet-ceph"Jenkins3-19/+0
2015-04-01puppet: wire in osfamily hieradata supportDan Prince1-7/+0
This updates all of the puppet roles to use an optional osfamily hieradata file which can be used to provide distro specific settings. Also, updates the controller role to make use of this new file for setting the rabbitmq package_provider parameter. Change-Id: I46417db51b87b82bf276dfcef5647a90c37fb07d
2015-04-01Merge "Set all RabbitMQ queues as mirrored when clustered"Jenkins1-0/+9
2015-04-01Merge "Ensure all Rabbit params are propagated to interested nodes."Jenkins1-29/+2
2015-03-31Merge "lb: Allow multiple backend"Jenkins1-1/+4
2015-03-31Merge "Refactor allNodesConfig"Jenkins1-1/+1
2015-03-31Merge "puppet+devtest: make compute nodes reachable"Jenkins1-0/+7
2015-03-31lb: Allow multiple backendYanis Guenane1-1/+4
Currently tripleo::loadbalancer allow a controller to have only itself as a backend for a service, no matter the number of controller nodes. This patch fixes that using all controller nodes available. Change-Id: Ic8fc022b84850c669b19d37da7f275d9c811e694 Depends-On: I2a46c250bc3325eef9c3128cac2ab45c88b1ae75
2015-03-28puppet: correct cinder enabled_backends formatDan Prince1-1/+1
This resolves a formatting issue with the Cinder enabled_backends config file setting. Previously we would potentially construct an array with an undef value at the end if iscsi was enabled but ceph was not (this is the case for our current CI job). When an array formatted like ['tripleo_iscsi', undef] is then passed to join() in puppet-cinder to construct a string it leaves us with an extra ',' on the end of the string. This causes problems in that cinder-volume loads an extra (system default) cinder volume process which is not expected. Because Fedora uses LIO as a default it was causing about half of our CI runs to fail if the tgtadm cinder-volume process wasn't being chosen by the scheduler. Closes-bug: #1437708 Change-Id: I3383012cb43792f334fdf789dc13147a3cb5ad63
2015-03-27Restructure Ceph/Puppet params to reflect changes in puppet-cephGiulio Fidente3-19/+0
A change [1] in puppet-ceph offers more flexibility but breaks backwards so we had to update our composition layer as well; we gain control of the cephx keyring in the template though. 1. Ie6adbd601388ab52c37037004bd0ceef9fc41942 Change-Id: Ia8196849afce2969daa608828cec81ebe3ac96e1
2015-03-26puppet+devtest: make compute nodes reachableJiri Stransky1-0/+7
Compute nodes run libvirt, which automatically creates a default network which has the same address space (192.168.122.*) as the libvirt default network on the host machine where devtest is running. This overlap causes that when a compute node wants to send a packet to the host machine (192.168.122.1) it gets incorrectly routed through the compute node's own virbr0 instead of br-ex. The current solution does not seem to be enough because libvirt gets started and creates the default network before Puppet is triggered on compute nodes. Making sure the libvirt default network is destroyed on the compute node fixes the issue. We don't have any puppet modules in OPM that would deal with libvirt networks and it's probably not worth exploring and adding one because of this small issue (i don't expect another use case of managing libvirt networks directly), so i'm using an exec with proper idempotency check. Change-Id: Icde12aa204ed1f7fa35b0525875ce07db34dc42c Closes-Bug: #1436822
2015-03-25Refactor allNodesConfigJiri Stransky1-1/+1
* Create hiera file 'all_nodes' instead of 'rabbit' -- we'll want allNodesConfig to create keys for more services (e.g. mongo_node_ips) and it's not necessary to create a separate hiera file for each. * Rename rabbit_nodes to mongo_node_names -- we'll have more node lists, some services will need hostnames, some services will need IPs, some might need both, so we shouldn't have ambiguity in the hiera key names. Change-Id: If80f9c9b2849ae893e1ab78f1c4d246a2468665c
2015-03-20Set all RabbitMQ queues as mirrored when clusteredGiulio Fidente1-0/+9
Purpose of this change is to enable on the server side the ha-mode policy for all queues when nodes are clustered. Change-Id: I16e3d375aabac9dbcdc198c71069086951e40fc0
2015-03-19Ensure all Rabbit params are propagated to interested nodes.Giulio Fidente1-29/+2
Change-Id: I1bb8ee15d361638d77c5df7f8c03561c34f4c88f
2015-03-19Merge "puppet: tidy up the Nova ceilo auth configs"Jenkins1-4/+1
2015-03-19Merge "puppet: tidy up the Nova glance API server config"Jenkins1-4/+1
2015-03-19Merge "puppet: tidy up the compute nova neutron config"Jenkins1-6/+1
2015-03-18Add support for Ceph as a Cinder and Nova backendYanis Guenane2-3/+54
This commit aims to add support for Ceph as a cinder and a nova backend. * Allows creation of Ceph pools from heat (Default: volumes, vms) * Creates the proper ceph user and inject the keys * Applies the proper configuration in cinder.conf and nova.conf * Enable the backend out of the box Co-Authored-By: Giulio Fidente <gfidente@redhat.com> Change-Id: Ic17d7a665de81a8bab5e34035abe90eda4bc889f
2015-03-18Merge "puppet/loadbalancer: use puppet-tripleo"Jenkins2-389/+9
2015-03-13puppet/loadbalancer: use puppet-tripleoEmilien Macchi2-389/+9
The loadbalancer Puppet code moved to puppet-tripleo (lightweight) composition layer. This patch aims to use it and refactor the loadbalancer.pp file. Co-Authored-By: Dan Prince <dprince@redhat.com> Change-Id: I1765ac9b6cb01cb64d5d28dad646674ddca859e9
2015-03-12puppet: tidy up the Nova ceilo auth configsDan Prince1-4/+1
Updates the puppet configuration for the Ceilometer auth agent so that we do the join conversions in the Heat templates and use only hiera for configuration of the ::ceilometer::agent::auth class. Change-Id: I932afafe21b2485a0581ac3910ac9d46161eee0d
2015-03-12puppet: tidy up the Nova glance API server configDan Prince1-4/+1
Updates the puppet configuration for the Nova glance configs so that we do the join conversions in the Heat templates and use only hiera for configuration of the ::nova class. Change-Id: Id12fb05470470558f1dccd45150bfce00a554466
2015-03-12puppet: tidy up the compute nova neutron configDan Prince1-6/+1
Updates the puppet configuration for the Nova neutron configs so that we do the join conversions in the Heat templates and use only hiera for configuration of the ::nova::network::neutron class. This updates the compute configuration to match what we now do on the controller as well. Change-Id: I2b352551777f64e0ceb119f48cc3b3ab1779f4d5
2015-03-12Use Cinder multibackend by defaultYanis Guenane2-2/+30
Currently Cinder iscsi backend is configured within the DEFAULT section. Since we aim to support multibackend, this commit puts the iscsi backend in its own section and enable it by default configuring it properly. Also adds a parameter which can be used to disable the default backend. Co-Authored-By: Giulio Fidente <gfidente@redhat.com> Change-Id: I05fb44b59829c0afa8a6588956a48320f2f65159
2015-03-12Merge "Puppet: Configure neutron_api_class on controller"Jenkins1-0/+1
2015-03-11Puppet: Configure neutron_api_class on controllerJiri Stransky1-0/+1
We're already configuring Neutron in Overcloud, but the controller is still configured to use the default Nova neutron_api_class for default configuration for networking, which means it used Nova Network and not Neutron. This causes some of the Nova API is_neutron checks to behave incorrectly. This patch updates the controller to use nova::network::neutron (like we already do on the overcloud_compute.pp role). As part of the change several of the compute specific hiera settings for the nova::network::neutron class have been moved to common.yaml. Change-Id: Id2d5a5a0aa1ca087de714880ef1ea98484b06849
2015-03-11Merge "puppet: add ordering for Neutron agents"Jenkins1-0/+5
2015-03-10puppet: add ordering for Neutron agentsEmilien Macchi1-0/+5
Before starting the Neutron agents, we need to make sure neutron-server is running so we don't have a race when starting the services. This patch adds some orchestration to do it. Change-Id: I24db069d6af1fadd302b0924f769db3f58f65685
2015-03-09Include cinder::glance into the ctrl manifestYanis Guenane1-0/+1
Include ::cinder::glance into the controller manifest to have proper (upstream) default value for cinder's glance related parameters. Change-Id: I9ac83b9e997d3c2502b08b642d4e41dba36ddf67
2015-03-06Puppet: define swift_store_auth_address in hieraEmilien Macchi1-3/+1
This patch updates the glance::backend::swift implementation to use only hiera variables instead of a mix of hiera, and inline class variables. Nothing was functionally wrong with the previous approach but now that we can compose more freely using the SoftwareDeployment defining all the variables in Hiera makes sense and is cleaner. Change-Id: I6d319841488d2ed94e088a5ac21e41dcd964ed1a Co-Authored-By: Dan Prince <dprince@redhat.com>