aboutsummaryrefslogtreecommitdiffstats
path: root/puppet/manifests/overcloud_controller_pacemaker.pp
AgeCommit message (Collapse)AuthorFilesLines
2015-11-17Merge "neutron: enable nova-event-callback by default"Jenkins1-0/+1
2015-11-17Merge "Pass ceph::pool arguments when calling class"Jenkins1-4/+2
2015-11-17neutron: enable nova-event-callback by defaultEmilien Macchi1-0/+1
* Add NovaApiVirtualIP string parameter. * Compute nova_url and nova_admin_auth_url parameters. * Configure in Hiera neutron::server::notifications::* parameters. * non-ha: include ::neutron::server::notifications * ha: include ::neutron::server::notifications and create orchestration * Set vif_plugging_is_fatal to True so we actually fail if Neutron is not able to create the VIF during Nova server creation workflow. Depends-On: I21dc10396e92906eab4651c318aa2ee62a8e03c7 Change-Id: I02e41f87404e0030d488476680af2f6d45af94ff
2015-11-16Merge "Bump further the stop/start timeout for pcmk/systemd services"Jenkins1-8/+8
2015-11-13Pass ceph::pool arguments when calling classGiulio Fidente1-4/+2
Pass the ceph::pool properties as arguments to the class call instead of setting them as class defaults. Ceph recommends max 32 PGs and min 4 PGs per OSD so this change also lowers the defaults to 32 which works with 1 OSD, suits well a scenario with 3 OSDs and is easy to customize in the static hiera if more than 8 OSDs are deployed. More info at: https://bugzilla.redhat.com/show_bug.cgi?id=1252546 Change-Id: Ifed11d1857900b2251dfdf69d6b6f168150e6330
2015-11-13Fix cinder error when CinderNfsMountOptions option is absentPierre Blanc1-1/+1
When I deploy director with NFS backend for cinder, sometimes I don't need nfs mount options. If I choose to omit this option, or if the option is defined to '', the deployment fails. This patch add just a default value for this option. Change-Id: Idf708aaecebd5c6db14f48ad2a53d6c2453be5ee Bugzilla: https://bugzilla.redhat.com/show_bug.cgi?id=1281870
2015-11-13Bump further the stop/start timeout for pcmk/systemd servicesGiulio Fidente1-8/+8
This bumps further up the stop/start timeout for the pcmk/systemd services so that it matches the 100s default set in future pcmk versions [1]. 1. https://github.com/ClusterLabs/pacemaker/commit/17d65e9f44061a4fa14a9cddd6edc403b2d6d2b3 Change-Id: I6fc18f1ad876c5a25723710a3b20d8ec9519dcba
2015-11-12Merge "Set default start/stop timeout for pcmk services to 95s"Jenkins1-8/+8
2015-11-11Merge "Allow customization of Ceph client user"Jenkins1-1/+1
2015-11-11Merge "Revert "Manage keystone initialization directly in t-h-t manifests""Jenkins1-21/+0
2015-11-10Merge "Allow customization of the Ceph pool names"Jenkins1-2/+2
2015-11-10Set default start/stop timeout for pcmk services to 95sGiulio Fidente1-8/+8
This change will increase the default start/stop timeout for all the non-ocf pcmk services to 95s to make sure it allows for at least 90s to the systemd script to complete the start/stop. More info at: https://bugzilla.redhat.com/show_bug.cgi?id=1275324 Change-Id: I04f691396a4118b456728a43d71d32ac9a556431
2015-11-05Allow customization of Ceph client userGiulio Fidente1-1/+1
Previously we enforced the Ceph user used by the OpenStack clients to be named 'openstack', this change allows for customization of such a name. Change-Id: Idef3e1ed4e8e21b645081869b8d6fad2329bdc60
2015-11-05Allow customization of the Ceph pool namesGiulio Fidente1-2/+2
This is useful in those scenarios were we want to use an external Ceph deployment with multiple overclouds. Change-Id: I1749d2a6547f6ce25843709e46a1447e8d42cfff
2015-11-05Make puppet manifests compliant with Puppet 4.xGael Chamoulaud1-285/+285
- https://docs.puppetlabs.com/puppet/3.8/reference/deprecated_language.html - Temporary disablement of the pupppet-lint autoload layout check failing for ringbuilder.pp. A fix for that will be part of an other patch. Change-Id: I495825641ab12e7c5789c1405649c356c5bb8051 Signed-off-by: Gael Chamoulaud <gchamoul@redhat.com>
2015-11-05Merge "Fix dependency for Ceilometer agent when using MySQL"Jenkins1-8/+17
2015-11-04Merge "Do not enforce start timeout to 90s for neutron-server"Jenkins1-1/+0
2015-11-04Revert "Manage keystone initialization directly in t-h-t manifests"Ben Nemec1-21/+0
This reverts commit 86d6c1ddc76bad423194e789ffb5474e4e12960e. This likely has an impact on upgrades, and since we don't have an upgrade CI job yet I'm concerned that we may have just broken ourselves. I would prefer to wait to merge this until the CI job is in place. Change-Id: Ib2366cb4b40471a28122f6e9955da9bdb31a53fb
2015-11-04Do not enforce start timeout to 90s for neutron-serverGiulio Fidente1-1/+0
The same timeout value is set for every pcmk service in [1] 1. https://github.com/openstack/tripleo-heat-templates/blob/master/puppet/manifests/overcloud_controller_pacemaker.pp#L81 Change-Id: I253f6cbc5ccdbf8c46cc537ff9600f201aae540a
2015-11-03Manage keystone initialization directly in t-h-t manifestsYanis Guenane1-0/+21
This is the second change of a servies of two, it creates the user, user_role, service and endpoint for: * glance * nova * neutron * cinder * horizon * swift * ceilometer * heat Change-Id: I50e792d98a2ba516ff498c58ad402f463c5f7e76
2015-11-03Create keystone roles and admin user from t-h-t manifestsYanis Guenane1-1/+20
Currently keystone initialization happens via os-cloud-config [1]. This commit moves some of that directly into the manifests. This is the first in a series of two changes to migrate it entirely into t-h-t. This change focus on implementing what keystone.initialize() was doing on the tripleoclient [2], creates the admin tenant, user and roles. It also creates the keystone endpoint itself. 1. https://github.com/openstack/os-cloud-config/blob/master/os_cloud_config/keystone.py#L128-L158 2. https://github.com/openstack/python-tripleoclient/blob/master/tripleoclient/v1/overcloud_deploy.py#L462-L527 Change-Id: I98555b707ff9b91c6e218de5dca68106ea05c8ea Depends-On: Ia4b3244f114dcff746ab89d355ad4933f8fdbddf
2015-11-03Fix dependency for Ceilometer agent when using MySQLJuan Antonio Osorio Robles1-8/+17
In HA, when using MySQL as a backend for Ceilometer, the dependencies set for the Ceilometer central agent depended always on MongoDB; Which should only be the case if MongoDB is set as a backend. Change-Id: I6fecfe0564b13e9352313c5a3492505b44d12eaa
2015-11-02Merge "Ensure memcached resource uses interleave=true"Jenkins1-1/+1
2015-11-02Merge "Support NFS backend for Glance (via Pacemaker)"Jenkins1-0/+10
2015-10-19Support NFS backend for Glance (via Pacemaker)Jiri Stransky1-0/+10
Adds support for NFS backend in Glance by allowing the storage directory for the 'file' backend to be a mount managed by Pacemaker. Default behavior is unchanged. Since the Pacemaker-related parameters are not exposed on top level, change storage-environment.yaml to use parameter_defaults instead of parameters. Depends on a Heat fix for environment file's parameter_defaults to work well with JSONs and comma delimited lists (see Depends-On). Change-Id: I6e7e2eaf6919b955650c0b32e1629a4067602c89 Depends-On: I85b13a79dbc97a77e20c0d5df8eaf05b3000815e
2015-10-19Ensure memcached resource uses interleave=trueGiulio Fidente1-1/+1
Without interleave Pacemaker will consider memcached as a single unit on all nodes so it will restart dependencies on every node in case of failures. More info at: https://bugzilla.redhat.com/show_bug.cgi?id=1262425 Change-Id: I284cbf33c8453a2abefc862f3d2e0b7f11ff4c3d
2015-10-15Sync httpd vhost settings in between pcmk and non-pcmk scenariosGiulio Fidente1-6/+0
Moves the vhost_params out of the manifest and into static hiera; also removes unneeded server_alias parameter as that matched the vhost servername anyway. Change-Id: I4b5971b23ef3be9529a59075fa93ccc64af75b9c
2015-10-12Merge "Ensure httpd is not enabled by puppet on system boot"Jenkins1-1/+4
2015-10-09Ensure httpd is not enabled by puppet on system bootGiulio Fidente1-1/+4
The httpd daemon will be started and managed by Pacemaker, it should not be enabled by puppet. Ideally, it shouldn't be started either but it seems it isn't possible with horizon and apache mod_wsgi [1]. 1. https://bugzilla.redhat.com/show_bug.cgi?id=1247547 Change-Id: I8a1b23c4ea27ac86385314f6cfde8c49d0879969 Co-Authored-By: marios andreou (marios@redhat.com)
2015-10-09Merge "Fix MariaDB account removal race condition"Jenkins1-6/+7
2015-10-09Merge "Rework pacemaker constraints from ovs and netns cleanup agents"Jenkins1-18/+22
2015-10-08Merge "puppet: allow to configure any ceilometer param with Hiera"Jenkins1-0/+1
2015-10-07Merge "Enable Cisco N1KV driver"Jenkins1-0/+44
2015-10-06Rework pacemaker constraints from ovs and netns cleanup agentsmarios1-18/+22
In the current neutron-* services constraints chain, the ovs and netns cleanup services are re-run after a neutron-server restart. As discussed at [1] this may not be desirable leaving some neutron services down and any tenant routers without IP. This review introduces a second constraints chain so we now have: neutron-server-->openvswitch-->dhcp-->l3-->metadata and ovs-cleanup-->netns-cleanup-->openvswitch Instead of a single chain like neutron-server-->ovs-cleanup-->netns-cleanup-->openvswitch--> dhcp-->l3-->metadata [1] https://bugzilla.redhat.com/show_bug.cgi?id=1266910#c12 Related-Bug: 1501378 Change-Id: I4096704257aff74ff5bd37d8d01d8a776c6c6a76
2015-10-05Fix MariaDB account removal race conditionJiri Stransky1-6/+7
The removal of default MariaDB accounts was being triggered roughly at the same time on all controllers, causing a race condition -- multiple nodes found an account present and attempted deletion, but then only one succeeded with the deletion, the others failed. HA controller deletes the accounts only on bootstrap node now, which fixes the issue. Change-Id: Ieacd10a6ce26da50f6a37eaa3221d866c24353fa
2015-10-02puppet: allow to configure any ceilometer param with HieraEmilien Macchi1-0/+1
By including ::ceilometer::config on controller & compute, we allow anyone to trick ceilometer.conf with any parameter, using Hiera. Change-Id: Ie6698d5e6900ecaaf7f19ed79e9c44b39ced0559
2015-10-01Write package names out to flat filesDan Prince1-0/+3
This patch updates all of the overcloud manifests so that we write out flat files containing lists of the Puppet packages which were managed by each manifest. The flat files all get written to /var/lib/puppet-tripleo/installed-packages/ where they can be easily parsed by external tools. Example format from the flat files looks like (for the controller step 1): cat /var/lib/puppet-tripleo/installed-packages/overcloud_controller1 keepalived haproxy Depends-On: If3e03b1983fed47082fac8ce63f975557dbc503c Change-Id: Ia324a08711796aa664f9c0273a051f4f2e3e92c9
2015-09-30Enable Cisco N1KV driverShiva Prasad Rao1-0/+44
This enables support for the Cisco N1kv driver for the ML2 plugin. It also configures the Nexus 1000v switch. Co-Authored-By: Steven Hillman <sthillma@cisco.com> Depends-On: I02dda0685c7df9013693db5eeacb2f47745d05b5 Depends-On: I3f14cdce9b9bf278aa9b107b2d313e1e82a20709 Change-Id: Idf23ed11a53509c00aa5fea4c87a515f42ad744f
2015-09-23Merge "Add redis ordering with ceilometer-central"Jenkins1-0/+16
2015-09-23Merge "Re-add constranits between nova-api and nova-novncproxy"Jenkins1-18/+16
2015-09-17Merge "Big Switch Neutron ML2 plugin integration"Jenkins1-0/+10
2015-09-16Merge "Set pacemaker default resource-stickiness"Jenkins1-0/+2
2015-09-16Big Switch Neutron ML2 plugin integrationJiri Stransky1-0/+10
Add support for Big Switch Neutron ML2 plugin. Makes sure that the package is present and sets up the [restproxy] section in ml2_conf.ini. This also adds support for setting the ovs_use_veth option in l3_agent.ini. There is no support for this in puppet-neutron l3 class and it probably doesn't make sense adding it there, because this setting isn't relevant for all l3 agent drivers, it's specific to OVSInterfaceDriver. The ovs_use_veth option is also added to dhcp_agent.ini. Change-Id: I99635e25b2099dacce68154fe14693d6f06ac19f
2015-09-16Merge "Enable Cisco Nexus and UCSM plugins"Jenkins1-0/+8
2015-09-16Merge "Add 4 pacemaker constraints for keystone"Jenkins1-0/+37
2015-09-15Enable Cisco Nexus and UCSM pluginsRobert Pothier1-0/+8
This enables support for the Cisco UCS Manager and Cisco Nexus plugins Change-Id: I1bc28a4768d5d6857a0504ca1f77dd71259570b8
2015-09-15Merge "Consume the NeutronMechanismDrivers from the hiera data"Jenkins1-0/+1
2015-09-10Merge "Set the nova scheduler ram_allocation_ration to 1.0"Jenkins1-0/+1
2015-09-07Don't add 'host' parameter to cinder.confDerek Higgins1-4/+0
As of I54a75652efd5e91464b84adf84004400b343c3a5 for rdb this is being done by the cinder puppet module. Change-Id: I109e139fcbb859a0d9ed99054656be94975d33b5
2015-09-02Add redis ordering with ceilometer-centralJiri Stransky1-0/+16
This is another missing constraint. The `require-all=false` part is good to have, otherwise Ceilometer (and transitively Heat) would switch to A/P mode. However, at the moment `require-all=false` isn't a recognized parameter on Fedora, hence the logic fork based on $::operatingsystem. Change-Id: I2657087192a05b2d8f0ab04ec60631d35331bf6c