summaryrefslogtreecommitdiffstats
path: root/playbooks/common
Commit message (Collapse)AuthorAgeFilesLines
* Move etcd backup from pre-upgrade to upgrade itself.Devan Goodwin2016-09-292-87/+86
|
* Allow a couple retries when unscheduling/rescheduling nodes in upgrade.Devan Goodwin2016-09-291-0/+12
| | | | | | | | This can fail with a transient "object has been modified" error asking you to re-try your changes on the latest version of the object. Allow up to three retries to see if we can get the change to take effect.
* Skip the docker role in early upgrade stages.Devan Goodwin2016-09-292-5/+6
| | | | | | | | This improves the situation further and prevents configuration changes from accidentally triggering docker restarts, before we've evacuated nodes. Now in two places, we skip the role entirely, instead of previous implementation which only skipped upgrading the installed version. (which did not catch config issues)
* Allow filtering nodes to upgrade by label.Devan Goodwin2016-09-297-13/+55
|
* Allow customizing node upgrade serial value.Devan Goodwin2016-09-291-1/+3
|
* Split upgrade for control plane/nodes.Devan Goodwin2016-09-2910-103/+89
|
* Verify masters are upgraded before proceeding with node only upgrade.Devan Goodwin2016-09-281-3/+0
|
* Attempt to tease apart pre upgrade for masters/nodes.Devan Goodwin2016-09-288-305/+251
|
* Split upgrade entry points into control plane/node.Devan Goodwin2016-09-282-1/+49
|
* Reunite upgrade reconciliation gating with the play it gates on.Devan Goodwin2016-09-281-17/+18
|
* Drop atomic-enterprise as a valid deployment type in upgrade.Devan Goodwin2016-09-281-2/+2
|
* Stop guarding against pacemaker in upgrade, no longer necessary.Devan Goodwin2016-09-281-8/+0
|
* Support openshift_upgrade_dry_run=true for pre-upgrade checks only.Devan Goodwin2016-09-281-0/+7
|
* Fix OpenStack cloud providerLénaïc Huard2016-09-262-8/+8
|
* Merge pull request #2295 from dgoodwin/override-docker-upgradeScott Dodson2016-09-151-1/+1
|\ | | | | Allow overriding the Docker 1.10 requirement for upgrade.
| * Allow overriding the Docker 1.10 requirement for upgrade.Devan Goodwin2016-08-151-1/+1
| | | | | | | | Respect an explicit docker_version, and the use of docker_upgrade=False.
* | Merge pull request #2449 from abutcher/service-signer-upgradeScott Dodson2016-09-143-1/+102
|\ \ | | | | | | [upgrade] Create/configure service signer cert when missing.
| * | [upgrade] Create/configure service signer cert when missing.Andrew Butcher2016-09-143-1/+102
| | |
* | | [redeploy-certificates] Set default value for ↵Andrew Butcher2016-09-141-0/+8
| | | | | | | | | | | | openshift_master_default_subdomain as workaround.
* | | [redeploy-certificates] Correct etcd service name.Andrew Butcher2016-09-141-1/+3
|/ /
* | Merge pull request #2425 from dgoodwin/flush-handlersScott Dodson2016-09-091-0/+2
|\ \ | | | | | | Flush handlers before marking a node schedulable after upgrade.
| * | Flush handlers before marking a node schedulable after upgrade.Devan Goodwin2016-09-091-0/+2
| | | | | | | | | | | | | | | | | | Handlers normally only trigger at the end of the play, but in this case we just set our node schedulable again resulting in it immediately getting taken down again.
* | | Iterate over node inventory hostnames instead of openshift.common.hostname ↵Andrew Butcher2016-09-081-3/+1
|/ / | | | | | | within openshift_manage_node role.
* | Don't loop over hostvars when setting node schedulability.Andrew Butcher2016-09-071-1/+0
| |
* | Merge pull request #2409 from abutcher/secure-registryScott Dodson2016-09-041-1/+83
|\ \ | | | | | | Secure registry for atomic registry deployment
| * | Secure registry for atomic registry deployment (deployment_subtype=registry).Andrew Butcher2016-09-021-1/+83
| | |
* | | Record schedulability of node prior to upgrade and re-set it to thatScott Dodson2016-09-021-1/+13
| | | | | | | | | | | | | | | | | | | | | Previously we were setting schedulability to the state defined in the inventory without regard to whether or not it was manually made schedulable or unschedulable. The right thing seems to be to record the state prior to upgrade and set it back.
* | | Fix ops/qps typoJordan Liggitt2016-09-022-3/+3
|/ /
* | Reload docker facts after upgrading dockerScott Dodson2016-08-301-0/+4
| |
* | Remove atomic check and cockpit.socketSamuel Munilla2016-08-262-2/+2
| |
* | Re-organize registry-console deployment.Andrew Butcher2016-08-262-2/+3
| |
* | Add support for Atomic Registry InstallsSamuel Munilla2016-08-261-1/+3
| | | | | | | | Add the Registry deployment subtype as an option in the quick installer.
* | Merge pull request #2052 from sdodson/metricsScott Dodson2016-08-253-1/+33
|\ \ | | | | | | Metrics improvements
| * | Deploy metrics after our routerScott Dodson2016-08-241-1/+1
| | | | | | | | | | | | | | | Metrics deployer now checks for route activation. As such we need a router before we install metrics.
| * | Pull in keynote demo changesScott Dodson2016-08-241-0/+6
| | |
| * | [tags] add some support for running a subset of config via tagsJason DeTiberus2016-08-242-0/+26
| | |
* | | Merge pull request #2337 from mkumatag/rm_dup_flannel_regAndrew Butcher2016-08-251-10/+0
|\ \ \ | | | | | | | | Remove duplicate flannel registration
| * | | Fix review commentsManjunath A Kumatagi2016-08-221-8/+2
| | | |
| * | | Remove duplicate flannel registrationManjunath A Kumatagi2016-08-211-4/+0
| | | |
* | | | Fix upgrade failure when master-config does not have pluginOrderOverride.Devan Goodwin2016-08-241-0/+1
| |/ / |/| |
* | | Merge pull request #2339 from dgoodwin/plugin-order-warningScott Dodson2016-08-231-0/+13
|\ \ \ | | | | | | | | Add warning at end of 3.3 upgrade if pluginOrderOverride is found.
| * | | Add warning at end of 3.3 upgrade if pluginOrderOverride is found.Devan Goodwin2016-08-221-0/+13
| | | |
* | | | Don't run node config upgrade hook if host is not a node.Devan Goodwin2016-08-231-1/+1
| | | |
* | | | Fix the “node on master” featureLénaïc Huard2016-08-191-1/+1
| |/ / |/| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | What we want to do is to add the master as a node if: * `g_nodeonmaster` is set to true, and * we are not in the case where we want to add new nodes. The second test was done by only checking whether `g_new_node_hosts` was defined. This was wrong because, in all cloud-provider setups, this variable was set with the default value of “empty list” (`[]`). The test has been changed to use the `bool` filter so that it correctly evaluates to false (and hence, effectively add the master as a node) when `g_new_node_hosts` is the empty list.
* | | Reconcile roles after master upgrade, but before nodes.Devan Goodwin2016-08-161-43/+46
|/ / | | | | | | | | | | | | | | | | | | Prevents the network egress bug causing node restart to fail during 3.3 upgrade. (even though a separate fix is incoming for this) Only catch is preventing the openshift_cli role, which requires docker, from triggering a potential upgrade, which we still don't want at this point. To avoid we use the same variable to protect docker installed version as we use in pre.yml.
* | Merge pull request #2220 from jkhelil/fix_key_error_when_ssh_failingScott Dodson2016-08-161-1/+1
|\ \ | |/ |/| fixing openshift key error in case of node failure during run (ssh is…
| * fixing openshift key error in case of node failure during run (ssh issue)jawed2016-08-091-1/+1
| |
* | Reference tmpdir from first master hostvars when evacuating nodes.Andrew Butcher2016-08-111-4/+4
| |
* | Merge pull request #2285 from dgoodwin/nuke-images-2Scott Dodson2016-08-114-20/+59
|\ \ | | | | | | Improvements for Docker 1.10+ Upgrade Image Nuking
| * | Improvements for Docker 1.10+ upgrade image nuking.Devan Goodwin2016-08-114-20/+55
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | In a parallel step prior to real upgrade tasks, clear out all unused Docker images on all hosts. This should be relatively safe to interrupt as no real upgrade steps have taken place. Once into actual upgrade, we again clear all images only this time with force, and after stopping and removing all containers. Both rmi commands use a new and hopefully less error prone command to do the removal, this should avoid missed orphans as we were hitting before. Added some logging around the current image count before and after this step, most of them are only printed if we're crossing the 1.10 boundary but one does not, just for additional information in your ansible log.