added upstart jobs, upload util library to HDFS, fix tracking log permissions

TODO.txt

@@ -13,8 +13,7 @@
 
 * document instructions
     * vagrant up
-    * as vagrant user run "/edx/app/edx_ansible/venvs/edx_ansible/bin/ansible-playbook -i localhost, -c loc
-al analytics_single.yml"
+    * as vagrant user run "/edx/app/edx_ansible/venvs/edx_ansible/bin/ansible-playbook -i localhost, -c local analytics_single.yml"
     * sudo su hadoop
         * edx-analytics-hadoop.sh start
 

playbooks/roles/analytics_pipeline/defaults/main.yml

@@ -24,12 +24,18 @@ ANALYTICS_PIPELINE_INPUT_DATABASE:
   port: 3306
 
 ANALYTICS_PIPELINE_CONFIG_DIR: "{{ COMMON_CFG_DIR }}/edx-analytics-pipeline"
+ANALYTICS_PIPELINE_HDFS_DATA_DIR: "hdfs://localhost:9000/data/"
 
 #
 # vars are namespaced with the module name.
 #
 analytics_pipeline_role_name: analytics_pipeline
 
+analytics_pipeline_util_library:
+  path: /tmp/edx-analytics-hadoop-util
+  repo: https://github.com/mulby/edx-analytics-hadoop-util
+  version: master
+
 #
 # OS packages
 #

playbooks/roles/analytics_pipeline/meta/main.yml

@@ -20,6 +20,7 @@
 # }
 
 dependencies:
+  - common
   - hadoop_master
   - hive
   - sqoop

playbooks/roles/analytics_pipeline/tasks/main.yml

@@ -39,8 +39,61 @@
     dest={{ COMMON_CFG_DIR }}/edx-analytics-pipeline/input.json
     mode=0644 owner={{ hadoop_common_user }} group={{ hadoop_common_group }}
 
+- name: util library source checked out
+  git: >
+    dest={{ analytics_pipeline_util_library.path }} repo={{ analytics_pipeline_util_library.repo }}
+    version={{ analytics_pipeline_util_library.version }}
+
+- name: lib directory created
+  file: >
+    path={{ HADOOP_COMMON_USER_HOME }}/lib
+    owner={{ hadoop_common_user }} group={{ hadoop_common_group }} state=directory
+
+- name: check if the util library needs to be built
+  stat: >
+    path={{ HADOOP_COMMON_USER_HOME }}/lib/edx-analytics-hadoop-util.jar
+  register: util_lib_built
+
+- name: util library built
+  shell: >
+    chdir={{ analytics_pipeline_util_library.path }}
+    {{ hadoop_common_java_home }}/bin/javac -cp `{{ HADOOP_COMMON_HOME }}/bin/hadoop classpath` org/edx/hadoop/input/ManifestTextInputFormat.java &&
+    {{ hadoop_common_java_home }}/bin/jar cf {{ HADOOP_COMMON_USER_HOME }}/lib/edx-analytics-hadoop-util.jar org/edx/hadoop/input/ManifestTextInputFormat.class &&
+    chown {{ hadoop_common_user }}:{{ hadoop_common_group }} {{ HADOOP_COMMON_USER_HOME }}/lib/edx-analytics-hadoop-util.jar
+  when: not util_lib_built.stat.exists
+
+- name: ensure hdfs services are started
+  service: >
+    name=hdfs
+    state=started
+
+- name: ensure map reduce services are started
+  service: >
+    name=yarn
+    state=started
+
+- name: ensure package dir exists in HDFS
+  shell: >
+    . {{ HADOOP_COMMON_CONF_DIR }}/hadoop-env.sh && hdfs dfs -mkdir -p /edx-analytics-pipeline/packages/
+  sudo_user: "{{ hadoop_common_user }}"
+
+- name: ensure util library is in HDFS
+  shell: >
+    . {{ HADOOP_COMMON_CONF_DIR }}/hadoop-env.sh && hdfs dfs -put -f {{ HADOOP_COMMON_USER_HOME }}/lib/edx-analytics-hadoop-util.jar /edx-analytics-pipeline/packages/
+  sudo_user: "{{ hadoop_common_user }}"
+
+- name: ensure the data directory exists
+  shell: >
+    . {{ HADOOP_COMMON_CONF_DIR }}/hadoop-env.sh && hdfs dfs -mkdir -p {{ ANALYTICS_PIPELINE_HDFS_DATA_DIR }}
+  sudo_user: "{{ hadoop_common_user }}"
+
+- name: ensure tracking log file can be read
+  file: >
+    path={{ COMMON_LOG_DIR }}/tracking/tracking.log
+    mode=0644
+
 - name: cron job syncs tracking log file to hdfs
   cron: >
     user={{ hadoop_common_user }}
     name="Sync tracking log to HDFS"
-    job="{{ HADOOP_COMMON_HOME }}/bin/hdfs dfs -put -f /edx/var/log/tracking/tracking.log hdfs://localhost:9000/data/tracking.log"
+    job="{{ HADOOP_COMMON_HOME }}/bin/hdfs dfs -put -f {{ COMMON_LOG_DIR }}/tracking.log {{ ANALYTICS_PIPELINE_HDFS_DATA_DIR }}/tracking.log"

playbooks/roles/hadoop_common/defaults/main.yml

@@ -42,10 +42,6 @@ hadoop_common_native_dist:
   url: "https://github.com/apache/hadoop-common/archive/release-{{ HADOOP_COMMON_VERSION }}.tar.gz"
   sha256sum: a8e1b49d4e891255d465e9449346ac7fb259bb35dce07d9f0df3b46fac3e9bd0
 hadoop_common_java_home: /usr/lib/jvm/java-7-oracle
-hadoop_common_util_library:
-  path: /tmp/edx-analytics-hadoop-util
-  repo: https://github.com/mulby/edx-analytics-hadoop-util
-  version: master
 hadoop_common_env: "{{ HADOOP_COMMON_USER_HOME }}/.hadoop_env"
 
 #

playbooks/roles/hadoop_common/tasks/main.yml

@@ -91,6 +91,15 @@
     - hdfs-site.xml
     - yarn-site.xml
 
+- name: upstart scripts installed
+  template: >
+    src={{ item }}.j2
+    dest=/etc/init/{{ item }}
+    mode=0640 owner=root group=root
+  with_items:
+    - hdfs.conf
+    - yarn.conf
+
 - name: hadoop env file exists
   file: >
     path={{ hadoop_common_env }} state=touch
@@ -180,38 +189,8 @@
     owner={{ hadoop_common_user }} group={{ hadoop_common_group }} state=touch
   when: not native_libs_built.stat.exists
 
-- name: util library source checked out
-  git: >
-    dest={{ hadoop_common_util_library.path }} repo={{ hadoop_common_util_library.repo }}
-    version={{ hadoop_common_util_library.version }}
-
-- name: lib directory created
-  file: >
-    path={{ HADOOP_COMMON_USER_HOME }}/lib
-    owner={{ hadoop_common_user }} group={{ hadoop_common_group }} state=directory
-
-- name: check if the util library needs to be built
-  stat: path={{ HADOOP_COMMON_USER_HOME }}/lib/edx-analytics-hadoop-util.jar
-  register: util_lib_built
-
-- name: util library built
-  shell: >
-    chdir={{ hadoop_common_util_library.path }}
-    {{ hadoop_common_java_home }}/bin/javac -cp `{{ HADOOP_COMMON_HOME }}/bin/hadoop classpath` org/edx/hadoop/input/ManifestTextInputFormat.java &&
-    {{ hadoop_common_java_home }}/bin/jar cf {{ HADOOP_COMMON_USER_HOME }}/lib/edx-analytics-hadoop-util.jar org/edx/hadoop/input/ManifestTextInputFormat.class &&
-    chown {{ hadoop_common_user }}:{{ hadoop_common_group }} {{ HADOOP_COMMON_USER_HOME }}/lib/edx-analytics-hadoop-util.jar
-  when: not util_lib_built.stat.exists
-
 - name: service directory exists
   file: >
     path={{ HADOOP_COMMON_SERVICES_DIR }}
     mode=0750 owner={{ hadoop_common_user }} group={{ hadoop_common_group }}
     state=directory
-
-- name: management script installed
-  template: >
-    src={{ item }}.j2
-    dest={{ COMMON_BIN_DIR }}/{{ item }}
-    mode=0755 owner={{ hadoop_common_user }} group={{ hadoop_common_group }}
-  with_items:
-    - edx-analytics-hadoop.sh

playbooks/roles/hadoop_common/templates/edx-analytics-hadoop.sh.j2

-#!/usr/bin/env bash
-
-set -e
-
-usage="Usage: edx-analytics-hadoop.sh [start|stop|restart]"
-
-function start() {
-    start-dfs.sh
-    start-yarn.sh
-}
-
-function stop() {
-    stop-dfs.sh
-    stop-yarn.sh
-}
-
-function restart() {
-	stop
-	start
-}
-
-case ${1-} in
-  start)
-        start
-        ;;
-  stop)
-        stop
-        ;;
-  restart)
-        restart
-        ;;
-  *)
-        echo $usage
-        exit 1
-        ;;
-esac

playbooks/roles/hadoop_common/templates/hdfs-site.xml.j2

@@ -6,14 +6,15 @@
     <name>dfs.replication</name>
     <value>1</value>
   </property>
-  
+
   <property>
     <name>dfs.namenode.name.dir</name>
     <value>file:{{ HADOOP_COMMON_DATA }}/namenode</value>
   </property>
-  
+
   <property>
-    <name>dfs.datanode.name.dir</name>
+    <name>dfs.datanode.data.dir</name>
     <value>file:{{ HADOOP_COMMON_DATA }}/datanode</value>
   </property>
+
 </configuration>
\ No newline at end of file

playbooks/roles/hadoop_common/templates/hdfs.conf.j2

+description     "hdfs"
+
+start on starting yarn
+stop on stopping yarn
+
+setuid {{ hadoop_common_user }}
+
+pre-start script
+  . {{ HADOOP_COMMON_CONF_DIR }}/hadoop-env.sh
+  start-dfs.sh
+end script
+
+post-stop script
+  . {{ HADOOP_COMMON_CONF_DIR }}/hadoop-env.sh
+  stop-dfs.sh
+end script

playbooks/roles/hadoop_common/templates/yarn.conf.j2

+description     "yarn"
+
+start on runlevel [2345]
+stop on runlevel [!2345]
+
+setuid {{ hadoop_common_user }}
+
+pre-start script
+  . {{ HADOOP_COMMON_CONF_DIR }}/hadoop-env.sh
+  start-yarn.sh
+end script
+
+post-stop script
+  . {{ HADOOP_COMMON_CONF_DIR }}/hadoop-env.sh
+  stop-yarn.sh
+end script

playbooks/roles/insights/defaults/main.yml

@@ -72,7 +72,7 @@ INSIGHTS_CONFIG:
   SECRET_KEY: '{{ INSIGHTS_SECRET_KEY }}'
   DATA_API_URL: '{{ ANALYTICS_API_ENDPOINT }}'
   DATA_API_AUTH_TOKEN: '{{ INSIGHTS_DATA_API_AUTH_TOKEN }}'
-  SOCIAL_AUTH_REDIRECT_IS_HTTPS: '{{ INSIGHTS_SOCIAL_AUTH_REDIRECT_IS_HTTPS | default(true) }}'
+  SOCIAL_AUTH_REDIRECT_IS_HTTPS: '{{ INSIGHTS_SOCIAL_AUTH_REDIRECT_IS_HTTPS | default(true) | bool }}'
   SOCIAL_AUTH_EDX_OIDC_KEY: '{{ INSIGHTS_OAUTH2_KEY }}'
   SOCIAL_AUTH_EDX_OIDC_SECRET: '{{ INSIGHTS_OAUTH2_SECRET }}'
   SOCIAL_AUTH_EDX_OIDC_URL_ROOT: '{{ INSIGHTS_OAUTH2_URL_ROOT }}'