Merge branch 'master' of github.com:openshift/openshift-ansible into variable_with_no_default_2576

8 files changed, 54 insertions, 23 deletions

diff --git a/.tito/packages/openshift-ansible b/.tito/packages/openshift-ansible
index da02d3f12..4daa2cb96 100644
--- a/.tito/packages/openshift-ansible
+++ b/.tito/packages/openshift-ansible
@@ -1 +1 @@
-3.4.5-1 ./
+3.4.6-1 ./
diff --git a/inventory/byo/hosts.origin.example b/inventory/byo/hosts.origin.example
index 706dd8080..0f0e223ce 100644
--- a/inventory/byo/hosts.origin.example
+++ b/inventory/byo/hosts.origin.example
@@ -599,7 +599,7 @@ openshift_master_identity_providers=[{'name': 'htpasswd_auth', 'login': 'true',
 #openshift_builddefaults_git_http_proxy=http://USER:PASSWORD@HOST:PORT
 #openshift_builddefaults_git_https_proxy=https://USER:PASSWORD@HOST:PORT
 # Or you may optionally define your own serialized as json
-#openshift_builddefaults_json='{"BuildDefaults":{"configuration":{"kind":"BuildDefaultsConfig","apiVersion":"v1","gitHTTPSProxy":"http://proxy.example.com.redhat.com:3128","gitHTTPProxy":"http://proxy.example.com.redhat.com:3128","env":[{"name":"HTTP_PROXY","value":"http://proxy.example.com.redhat.com:3128"},{"name":"NO_PROXY","value":"ose3-master.example.com"}]}}'
+#openshift_builddefaults_json='{"BuildDefaults":{"configuration":{"kind":"BuildDefaultsConfig","apiVersion":"v1","gitHTTPSProxy":"http://proxy.example.com.redhat.com:3128","gitHTTPProxy":"http://proxy.example.com.redhat.com:3128","env":[{"name":"HTTP_PROXY","value":"http://proxy.example.com.redhat.com:3128"},{"name":"HTTPS_PROXY","value":"http://proxy.example.com.redhat.com:3128"},{"name":"NO_PROXY","value":"ose3-master.example.com"}]}}'
 # masterConfig.volumeConfig.dynamicProvisioningEnabled, configurable as of 1.2/3.2, enabled by default
 #openshift_master_dynamic_provisioning_enabled=False
 
diff --git a/inventory/byo/hosts.ose.example b/inventory/byo/hosts.ose.example
index 09f1350b1..2e5b7564d 100644
--- a/inventory/byo/hosts.ose.example
+++ b/inventory/byo/hosts.ose.example
@@ -599,7 +599,7 @@ openshift_master_identity_providers=[{'name': 'htpasswd_auth', 'login': 'true',
 #openshift_builddefaults_git_http_proxy=http://USER:PASSWORD@HOST:PORT
 #openshift_builddefaults_git_https_proxy=https://USER:PASSWORD@HOST:PORT
 # Or you may optionally define your own serialized as json
-#openshift_builddefaults_json='{"BuildDefaults":{"configuration":{"kind":"BuildDefaultsConfig","apiVersion":"v1","gitHTTPSProxy":"http://proxy.example.com.redhat.com:3128","gitHTTPProxy":"http://proxy.example.com.redhat.com:3128","env":[{"name":"HTTP_PROXY","value":"http://proxy.example.com.redhat.com:3128"},{"name":"NO_PROXY","value":"ose3-master.example.com"}]}}'
+#openshift_builddefaults_json='{"BuildDefaults":{"configuration":{"kind":"BuildDefaultsConfig","apiVersion":"v1","gitHTTPSProxy":"http://proxy.example.com.redhat.com:3128","gitHTTPProxy":"http://proxy.example.com.redhat.com:3128","env":[{"name":"HTTP_PROXY","value":"http://proxy.example.com.redhat.com:3128"},{"name":"HTTPS_PROXY","value":"http://proxy.example.com.redhat.com:3128"},{"name":"NO_PROXY","value":"ose3-master.example.com"}]}}'
 # masterConfig.volumeConfig.dynamicProvisioningEnabled, configurable as of 1.2/3.2, enabled by default
 #openshift_master_dynamic_provisioning_enabled=False
 
diff --git a/openshift-ansible.spec b/openshift-ansible.spec
index f4e89f906..5aaa4c2bc 100644
--- a/openshift-ansible.spec
+++ b/openshift-ansible.spec
@@ -5,7 +5,7 @@
 }
 
 Name:           openshift-ansible
-Version:        3.4.5
+Version:        3.4.6
 Release:        1%{?dist}
 Summary:        Openshift and Atomic Enterprise Ansible
 License:        ASL 2.0
@@ -249,6 +249,12 @@ Atomic OpenShift Utilities includes
 
 
 %changelog
+* Mon Oct 10 2016 Troy Dawson <tdawson@redhat.com> 3.4.6-1
+- Retry failed master startup once (ironcladlou@gmail.com)
+- [logging] Fix openshift_hosted_logging_fluentd_nodeselector
+  (sdodson@redhat.com)
+- Changes for etcd servers (vishal.patil@nuagenetworks.net)
+
 * Fri Oct 07 2016 Scott Dodson <sdodson@redhat.com> 3.4.5-1
 - [a-o-i] -v disables quiet ansible config. (abutcher@redhat.com)
 
diff --git a/roles/openshift_facts/library/openshift_facts.py b/roles/openshift_facts/library/openshift_facts.py
index 9ffd399bc..0ee018c5c 100755
--- a/roles/openshift_facts/library/openshift_facts.py
+++ b/roles/openshift_facts/library/openshift_facts.py
@@ -364,12 +364,15 @@ def normalize_openstack_facts(metadata, facts):
     facts['network']['ip'] = local_ipv4
     facts['network']['public_ip'] = metadata['ec2_compat']['public-ipv4']
 
-    # TODO: verify local hostname makes sense and is resolvable
-    facts['network']['hostname'] = metadata['hostname']
-
-    # TODO: verify that public hostname makes sense and is resolvable
-    pub_h = metadata['ec2_compat']['public-hostname']
-    facts['network']['public_hostname'] = pub_h
+    for f_var, h_var, ip_var in [('hostname',        'hostname',        'local-ipv4'),
+                                 ('public_hostname', 'public-hostname', 'public-ipv4')]:
+        try:
+            if socket.gethostbyname(metadata['ec2_compat'][h_var]) == metadata['ec2_compat'][ip_var]:
+                facts['network'][f_var] = metadata['ec2_compat'][h_var]
+            else:
+                facts['network'][f_var] = metadata['ec2_compat'][ip_var]
+        except socket.gaierror:
+            facts['network'][f_var] = metadata['ec2_compat'][ip_var]
 
     return facts
 
diff --git a/roles/openshift_hosted/tasks/registry/storage/persistent_volume.yml b/roles/openshift_hosted/tasks/registry/storage/persistent_volume.yml
index 60eefd71a..0172f5ca0 100644
--- a/roles/openshift_hosted/tasks/registry/storage/persistent_volume.yml
+++ b/roles/openshift_hosted/tasks/registry/storage/persistent_volume.yml
@@ -3,7 +3,11 @@
     registry_volume_claim: "{{ openshift.hosted.registry.storage.volume.name }}-claim"
 
 - name: Determine if volume is already attached to dc/docker-registry
-  command: "{{ openshift.common.client_binary }} get -o template dc/docker-registry --template=\\{\\{.spec.template.spec.volumes\\}\\} --output-version=v1"
+  command: >
+      {{ openshift.common.client_binary }}
+      --config={{ openshift_hosted_kubeconfig }}
+      --namespace={{ openshift.hosted.registry.namespace | default('default') }}
+      get -o template dc/docker-registry --template=\\{\\{.spec.template.spec.volumes\\}\\} --output-version=v1
   changed_when: false
   failed_when: false
   register: registry_volumes_output
@@ -13,7 +17,10 @@
 
 - name: Add volume to dc/docker-registry
   command: >
-    {{ openshift.common.client_binary }} volume dc/docker-registry
+    {{ openshift.common.client_binary }}
+    --config={{ openshift_hosted_kubeconfig }}
+    --namespace={{ openshift.hosted.registry.namespace | default('default') }}
+    volume dc/docker-registry
     --add --overwrite -t persistentVolumeClaim --claim-name={{ registry_volume_claim }}
     --name=registry-storage
   when: not volume_attached | bool
diff --git a/roles/openshift_master/tasks/main.yml b/roles/openshift_master/tasks/main.yml
index ce2f96723..1a59717c7 100644
--- a/roles/openshift_master/tasks/main.yml
+++ b/roles/openshift_master/tasks/main.yml
@@ -168,10 +168,21 @@
 - include: set_loopback_context.yml
   when: openshift.common.version_gte_3_2_or_1_2
 
+# TODO: Master startup can fail when ec2 transparently reallocates the block
+# storage, causing etcd writes to temporarily fail. Retry failures blindly just
+# once to allow time for this transient condition to to resolve and for systemd
+# to restart the master (which will eventually succeed).
+#
+# https://github.com/coreos/etcd/issues/3864
+# https://github.com/openshift/origin/issues/6065
+# https://github.com/openshift/origin/issues/6447
 - name: Start and enable master
   service: name={{ openshift.common.service_type }}-master enabled=yes state=started
   when: not openshift_master_ha | bool
   register: start_result
+  until: not start_result | failed
+  retries: 1
+  delay: 60
   notify: Verify API Server
 
 - name: Check for non-HA master service presence
@@ -202,6 +213,9 @@
     state: started
   when: openshift_master_ha | bool and openshift.master.cluster_method == 'native' and inventory_hostname == openshift_master_hosts[0]
   register: start_result
+  until: not start_result | failed
+  retries: 1
+  delay: 60
 
 - set_fact:
     master_api_service_status_changed: "{{ start_result | changed }}"
@@ -218,6 +232,9 @@
     state: started
   when: openshift_master_ha | bool and openshift.master.cluster_method == 'native' and inventory_hostname != openshift_master_hosts[0]
   register: start_result
+  until: not start_result | failed
+  retries: 1
+  delay: 60
 
 - set_fact:
     master_api_service_status_changed: "{{ start_result | changed }}"
@@ -251,6 +268,9 @@
     state: started
   when: openshift_master_ha | bool and openshift.master.cluster_method == 'native' and inventory_hostname == openshift_master_hosts[0]
   register: start_result
+  until: not start_result | failed
+  retries: 1
+  delay: 60
 
 - pause:
     seconds: 15
@@ -263,6 +283,9 @@
     state: started
   when: openshift_master_ha | bool and openshift.master.cluster_method == 'native' and inventory_hostname != openshift_master_hosts[0]
   register: start_result
+  until: not start_result | failed
+  retries: 1
+  delay: 60
 
 - set_fact:
     master_controllers_service_status_changed: "{{ start_result | changed }}"
diff --git a/roles/openshift_node/tasks/main.yml b/roles/openshift_node/tasks/main.yml
index be07bd2d3..8e9c9f511 100644
--- a/roles/openshift_node/tasks/main.yml
+++ b/roles/openshift_node/tasks/main.yml
@@ -152,17 +152,9 @@
 - name: Start and enable node
   service: name={{ openshift.common.service_type }}-node enabled=yes state=started
   register: node_start_result
-  ignore_errors: yes
-
-- name: Wait 30 seconds for docker initialization whenever node has failed
-  pause:
-    seconds: 30
-  when: node_start_result | failed
-
-- name: Start and enable node again
-  service: name={{ openshift.common.service_type }}-node enabled=yes state=started
-  register: node_start_result
-  when: node_start_result | failed
+  until: not node_start_result | failed
+  retries: 1
+  delay: 30
 
 - set_fact:
     node_service_status_changed: "{{ node_start_result | changed }}"