summaryrefslogtreecommitdiffstats
path: root/roles/etcd
Commit message (Collapse)AuthorAgeFilesLines
* Docker 1.10 UpgradeDevan Goodwin2016-06-031-6/+6
| | | | | | | | | | | | | | | | | Adds a separate playbook for Docker 1.10 upgrade that can be run standalone on a pre-existing 3.2 cluster. The upgrade will take each node out of rotation, and remove *all* containers and images on it, as this is reportedly faster and more storage efficient than performing the in place 1.10 upgrade. This process is integrated into the 3.1 to 3.2 upgrade process. Normal config playbooks now become 3.2 only, and require Docker 1.10. Users of older environments will have to use an appropriate openshift-ansible version. Config playbooks no longer are in the business of upgrading or downgrading docker.
* Merge pull request #1971 from dgoodwin/missing-services-fixJason DeTiberus2016-06-021-6/+6
|\ | | | | Fix error with stopping services that may not exist.
| * Fix error with stopping services that may not exist.Devan Goodwin2016-06-021-6/+6
| | | | | | | | | | | | Causes a hard failure due to missing etcd/atomic-openshift-master services. Instead, check that the services exist before ensuring they're stopped/disabled.
* | Move os_firewall_allow from defaults to role dependencies.Andrew Butcher2016-06-012-8/+7
|/
* Default os_firewall_use_firewalld to false in os_firewall and remove overrides.Andrew Butcher2016-05-021-1/+0
|
* Remove etcd_interface variable. Remove openshift_docker dependency from the ↵Andrew Butcher2016-04-202-2/+0
| | | | etcd role.
* Use openshift_hostname/openshift_ip values for etcd configuration and ↵Andrew Butcher2016-04-203-31/+30
| | | | certificates.
* The openshift_docker role must set the version facts for containerized installsBrenton Leanhardt2016-04-141-0/+1
| | | | | | QE found that for fresh installs we were basing the docker version facts of the images that could be pulled prior to configuring /etc/sysconfig/docker. This is an edge case but something we need to fix.
* use RestartSec to avoid default rate limit in systemdBrenton Leanhardt2016-04-121-0/+1
| | | | Lifted from https://github.com/openshift/origin/pull/8317
* Bug 1320829 - Ensure docker installed for factsJason DeTiberus2016-03-311-1/+2
| | | | | | - gather facts requiring docker only if docker is present and running - Update reference to etcd role in playbooks/common/openshift-etcd/config.yml to use openshift_etcd
* make docker service want ose containerized servicesSeth Jennings2016-03-281-1/+1
| | | | | | | | | | | | | | | | There are points where the docker service is restarted during the install. Sometimes the services that are PartOf docker.service do not get restarted when docker is restarted. https://bugzilla.redhat.com/show_bug.cgi?id=1318948 Systemd documentation recommends using "wants" as the perferred means of creating startup linkages between units. This patch makes the ose services wanted by the docker service rather than multi-user.target. This creates a downward link from the docker service to the ose containerized services.
* Only mask etcd service for containerized installls when it's installedScott Dodson2016-03-151-1/+5
|
* improve ordering of systemd unitsJason DeTiberus2016-03-142-0/+7
|
* Ensure is_containerized is cast as bool.Andrew Butcher2016-03-141-1/+1
|
* Bug 1313169 - Ansible installer tries to enable etcd_container service even ↵Brenton Leanhardt2016-03-021-1/+1
| | | | though containerized=false
* remove version requirement from etcd, shouldn't be needed anymoreAdam Miller2016-02-221-1/+1
|
* Always pull docker imagesScott Dodson2016-01-071-16/+0
|
* Fix typo in etcd service status fact.Andrew Butcher2016-01-041-1/+1
|
* Fix restart handlers.Andrew Butcher2015-12-262-2/+2
|
* Add some guards to wait for images to be pulled before moving onScott Dodson2015-12-161-0/+10
|
* pull docker images only if not already presentJason DeTiberus2015-12-151-3/+8
|
* fixesJason DeTiberus2015-12-151-10/+7
|
* Containerization work by @sdodsonScott Dodson2015-12-155-14/+63
|
* Initial containerization work from @ibottyTobias Florek2015-12-151-1/+1
| | | | copied from https://github.com/eparis/kubernetes-ansible/blob/17f98edd7ff53e649b43e26822b8fbc0be42b233/roles/common/tasks/main.yml
* Remove yum / dnf duplicationScott Dodson2015-12-091-6/+1
|
* Fedora changes:Adam Miller2015-12-022-1/+6
| | | | | | | | | | | | | - ansible bootstrap playbook for Fedora 23+ - add conditionals to handle yum vs dnf - add Fedora OpenShift COPR - update BYO host README for repo configs and fedora bootstrap Fix typo in etcd README, remove unnecessary parens in openshift_node main.yml rebase on master, update package cache refresh handler for yum vs dnf Fix typo in etcd README, remove unnecessary parens in openshift_node main.yml
* Fix etcd cert generation when etcd_interface is definedJason DeTiberus2015-11-026-14/+15
| | | | | | | | - Refactor certificate generation to properly accept overrides of etcd_interface per host and set the certificate SANS and peer URLs properly. - Add sanity checking to user-set values of etcd_interface to provide a better error message
* UpgradesScott Dodson2015-09-231-0/+1
|
* Require etcd 2.*Scott Dodson2015-08-131-1/+1
| | | | | | | Fixes #422 When etcd-2.1 is available in RHEL7 / Centos 7 we'll bumpt to that as it's considerably more stable with regard to WAL corruption and recovery.
* Reduce heartbeat frequency to 500ms to reduce etcd cpu loadScott Dodson2015-08-051-2/+2
| | | | | | | Per https://github.com/coreos/etcd/pull/3097 reduce heartbeat to 500ms until we can ensure etcd 2.1 is in use Reduces the impact of but doesn't fix BZ1250310
* Update dependencies to ensure repos are available before openshift_master_* ↵Jason DeTiberus2015-07-171-0/+1
| | | | and etcd_* roles are run
* Fix etcd default interfaceJason DeTiberus2015-07-171-1/+1
| | | | | - instead of hardcoding to eth0, use the interface associated with the ipv4 default route as the default instead.
* add pauses to avoid service restarts from interfering with initial startup ↵Jason DeTiberus2015-07-151-1/+3
| | | | bootstrapping
* Etcd role updates and playbook updatesJason DeTiberus2015-07-104-12/+57
| | | | | | | | | | | | | | | | | | | | | - fix firewall conflict issues with co-located etcd and openshift hosts - added os_firewall dependency to etcd role - updated etcd template to better handle clustered and non-clustered installs - added etcd_ca role - generates a self-signed cert to manage etcd certificates, since etcd peer certificates are required to be client and server certs and the openshift ca will only generate client or server certs (not one authorized for both). - renamed openshift_etcd_certs role to etcd_certificates and updated it to manage certificates generated from the CA managed by the etcd_ca role - remove hard coded etcd_port in openshift_facts - updates for the openshift-etcd common playbook - removed etcd and openshift-etcd playbooks from the byo playbooks directory - added a common playbook for setting etcd launch facts - added an openshift-etcd common service playbook - removed unused variables - fixed tests for embedded_{etcd,dns,kube} in openshift_master - removed old workaround for reloading systemd units
* Add etcd role that builds out basic etcd clusterScott Dodson2015-07-106-0/+149
| | | | | | - Add initial etcd role - Add etcd playbook to create etcd client certs - Hookup master to etcd
* Removed cloud.rb and it's associated libraries as it's no longer used. Also ↵Thomas Wiest2015-05-275-219/+0
| | | | removed etcd role.
* Cleanup empty role dirs/filesJason DeTiberus2015-02-242-4/+0
|
* Initial Commit. Sharing is caringKenny Woodson2014-09-167-0/+223