summaryrefslogtreecommitdiffstats
path: root/playbooks/gce
Commit message (Collapse)AuthorAgeFilesLines
* Fix GCE cluster creationLénaïc Huard2016-11-162-543/+1
| | | | | | | | | | | Attempting to create a GCE cluster when the `gce.ini` configuration file contains a non-default network leads to the following error: ``` TASK [Launch instance(s)] ****************************************************** fatal: [localhost]: FAILED! => {"changed": false, "failed": true, "msg": "Unexpected error attempting to create instance lenaic2-master-74f10, error: {'domain': 'global', 'message': \"Invalid value for field 'resource.networkInterfaces[0]': ''. Subnetwork should be specified for custom subnetmode network\", 'reason': 'invalid'}"} ``` The `subnetwork` parameter needs to be added and taken into account.
* Template with_items for upstream ansible-2.2 compat.Andrew Butcher2016-10-144-8/+8
|
* Fix GCE LaunchBrad Durrow2016-08-071-1/+1
| | | The Ansible GCE module (documentation here: http://docs.ansible.com/ansible/gce_module.html) requires a comma separated list when you pass an array here (even with a single element) the argument has square brackets around it and the instance doesn't get launched. Testing shows that joining with ', ' (comma space) works with one instance but breaks with two so I used ',' (comma no space).
* Resolve some deprecation warnings.Andrew Butcher2016-07-181-1/+1
|
* Various hosted component improvementsAndrew Butcher2016-07-071-2/+1
| | | | | | | | | | | | | * [openshift_projects] Add openshift_projects role * [openshift_hosted] hosted deployments use openshift_hosted_infra_selector if openshift_hosted_<component>_selector is not defined * [openshift_hosted] move openshift_projects, openshift_serviceaccounts and openshift_metrics to dependencies of openshift_hosted * [router] improve router deployment - add router option to force subdomain - add CA to router certificate options * [registry] move registry config into openshift_hosted role - additional registry fixes/tweaks - add s3 storage support for registry * [serviceaccount] fix up serviceaccount creation
* g_all_hosts with templated with_items causes errors with ansible 1.9.4 under ↵Jason DeTiberus2016-05-052-3/+3
| | | | some conditions
* gce fixesJason DeTiberus2016-05-054-6/+582
|
* fix up variable references remove "online" support from bin/clusterJason DeTiberus2016-04-291-5/+0
|
* Replace sudo with become when accessing deployment_vars.Andrew Butcher2016-04-276-7/+7
|
* Replace deprecated sudo with become.Andrew Butcher2016-04-256-9/+9
|
* Fix router selector fact migration and match multiple selectors when ↵Andrew Butcher2016-04-201-1/+1
| | | | counting nodes.
* Fix issue when there are no infra nodesLénaïc Huard2016-03-081-1/+1
| | | | This issue was already addressed by #1355 fc2f776 but it has been accidentally reverted by 6d3e176.
* Move additional master configuration into a separate master playbook.Andrew Butcher2016-02-161-2/+0
|
* Add new_masters to scaleup playbook.Andrew Butcher2016-02-161-1/+7
|
* Merge pull request #1369 from lhuard1A/gce_oo_optionThomas Wiest2016-02-151-4/+4
|\ | | | | Make the GCE image_name and the machine_type configurable from the CLI
| * Make the GCE image_name and the machine_type configurable from the CLILénaïc Huard2016-02-101-4/+4
| |
* | Merge pull request #1355 from lhuard1A/fix_when_no_infraThomas Wiest2016-02-151-2/+2
|\ \ | | | | | | Fix issue when there are no infra nodes
| * | Fix issue when there are no infra nodesLénaïc Huard2016-02-091-2/+2
| |/
* | Merge pull request #1357 from lhuard1A/pretty_print_gce_list_clusterThomas Wiest2016-02-151-2/+9
|\ \ | | | | | | Better structure the output of the list playbook
| * | Better structure the output of the list playbookLénaïc Huard2016-02-091-2/+9
| |/ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The list playbook listed the IPs of the VMs without logging their role like: ``` PLAY [List instance(s)] ******************************************************* TASK: [debug ] **************************************************************** ok: [lenaic-node-compute-cd81b] => { "msg": "public ip: 104.155.50.164 private ip:10.55.0.49" } ok: [lenaic-master-9e767] => { "msg": "public ip: 104.155.47.180 private ip:10.55.0.113" } ok: [lenaic-node-infra-ab7c8] => { "msg": "public ip: 104.155.11.34 private ip:10.55.0.131" } ``` The list playbook now prints the information in a more structured way like: ``` PLAY [List Hosts] ************************************************************* TASK: [debug ] **************************************************************** ok: [localhost] => { "msg": { "lenaic": { "master": { "default": [ { "name": "lenaic-master-9e767", "private IP": "10.55.0.113", "public IP": "104.155.47.180" } ] }, "node": { "compute": [ { "name": "lenaic-node-compute-cd81b", "private IP": "10.55.0.49", "public IP": "104.155.50.164" } ], "infra": [ { "name": "lenaic-node-infra-ab7c8", "private IP": "10.55.0.131", "public IP": "104.155.11.34" } ] } } } } ``` This change of the output of the list playbook was previously done for OpenStack and libvirt in 332aa8c (#461). This change makes the GCE output identical to OpenStack and libvirt’s one.
* | Merge pull request #1364 from codificat/openshift-enterpriseBrenton Leanhardt2016-02-101-1/+1
|\ \ | |/ |/| Add missing atomic- and openshift-enterprise
| * Add missing atomic- and openshift-enterprisePep Turró Mauri2016-02-091-1/+1
| | | | | | | | | | | | Some checks related to *enterprise deployments were still only looking for "enterprise" deployment_type. Update them to cover also atomic-enterprise and openshift-enterprise deployment types.
* | Merge pull request #1239 from detiber/bin_cluster_overridesThomas Wiest2016-02-031-0/+4
|\ \ | | | | | | Additional overrides for cloud provider playbooks
| * | Additional overrides for cloud provider playbooksJason DeTiberus2016-02-011-0/+4
| | | | | | | | | | | | | | | - sdn overrides - allow overrides for use_flannel and use_fluentd
* | | Merge pull request #1173 from abutcher/nfs-groupBrenton Leanhardt2016-02-031-0/+1
|\ \ \ | |_|/ |/| | Refactor storage options
| * | Refactor registry storage options.Andrew Butcher2016-02-011-0/+1
| | |
* | | Improve docs and consistency of setting the ssh_userJason DeTiberus2016-02-027-9/+9
| | |
* | | remove outdated commentsJason DeTiberus2016-02-022-5/+0
| | |
* | | add etcd hosts for gce playbooksJason DeTiberus2016-02-021-0/+11
| | |
* | | GCE cloud provider updatesJason DeTiberus2016-02-027-94/+49
| |/ |/| | | | | | | | | | | | | | | | | | | | | - fix overrides for gce machine type and gce machine image - Update default image for origin - Update default ssh user for origin and enterprise - Remove old commented out code - Remove wip and join_node playbooks - Added add_nodes playbook which will now allow for using bin/cluster to add additional nodes - Allow env override of ssh_user - improve list playbook
* | Multi-master fixes for provider playbooksJason DeTiberus2016-02-011-0/+1
|/ | | | | - set openshift_master_cluster_method=native for all cloud providers so bin/cluster will build the ha masters correctly
* infra_node fixesJason DeTiberus2016-02-011-0/+2
| | | | | | | | | | | | | | - openshift_master role update - infra_nodes was previously being set to num_infra, which is an integer value when using the cloud providers, added a new variable osm_infra_nodes that is expected to be a list of hosts - if openshift_infra_nodes is not already set, create it from the nodes that have the region=infra label. - Cloud provider config playbook updates - override openshift_router_selector for cloud providers to avoid using the default of 'region=infra' when deployment_type is not 'online' - Set openshift_infra_nodes to g_infra_host for cloud providers
* Update cluster_hosts.yml for cloud providersJason DeTiberus2016-01-191-16/+11
| | | | | | - Add g_infra_hosts (nodes with sub-type infra) - Add g_compute_hosts (nodes with sub-type compute) - Reduce duplication by re-using previously defined variables
* Remove default disable of SDN for GCEJason DeTiberus2016-01-153-6/+0
| | | | | | - Update openshift_facts to no longer default use_openshift_sdn for GCE installs - Remove workarounds in the GCE playbooks that were forcing SDN configuration
* Make bin/cluster able to spawn OSE 3.1 clustersLénaïc Huard2016-01-081-5/+10
|
* Fix error in byo cluster_hosts.ymlJason DeTiberus2016-01-041-0/+3
|
* Cleanup and fixes for cluster_id changeJason DeTiberus2016-01-044-48/+25
| | | | | | | | | - Move debug_level into vars.yml and byo inventory - change variables in cluster_hosts.yml to be g_* and update playbooks to use those values directly instead of setting them indirectly - added a new g_all_hosts entry in cluster_hosts to use in the update playbook instead of unioning all host types within the playbook - added a cluster_hosts.yml for the byo playbook
* Removing environment and env tags.Kenny Woodson2016-01-048-24/+44
|
* Create nfs host group with registry volume attachment.Andrew Butcher2015-12-151-0/+1
|
* Merge pull request #1028 from kwoodson/remove_env_host_typeKenny Woodson2015-12-146-15/+20
|\ | | | | Removing env-host-type in preparation of env and environment changes.
| * Updating env-host-type to host patternsKenny Woodson2015-12-116-15/+20
| |
* | Merge pull request #1000 from detiber/localhostFixThomas Wiest2015-12-108-0/+16
|\ \ | | | | | | Enforce connection: local and become: no on all localhost plays
| * | Enforce connection: local and become: no on all localhost playsJason DeTiberus2015-11-308-0/+16
| |/
* / Pass in and use first_master_ip as dnsIP for pre 3.1 nodes.Andrew Butcher2015-12-081-2/+0
|/
* Differentiate machine types on GCE (master and nodes)Romain Dossin2015-11-193-2/+9
|
* Fix lb group related errorsJason DeTiberus2015-11-051-0/+1
|
* Various HA changes for pacemaker and native methods.Andrew Butcher2015-11-051-1/+1
|
* Refactor common group evaluation to avoid duplicationJason DeTiberus2015-11-041-3/+3
|
* Fixed GCE playbooks so that they're more like the AWS playbooks.Thomas Wiest2015-10-143-18/+35
| | | | | | | | Namely the GCE playbooks now: - Create infra nodes - Correctly label nodes in OpenShift - Setup masters as nodes as well (needed for sdn) - Removed set_infra_launch_facts_tasks.yml as it's not used anymore.
* Adding second param. true to many default filtersChengcheng Mu2015-10-093-6/+6
|