rhel7-provision-solr.yml

#
# Install Solr
#
- name: Install Solr w/tomcat.
  remote_user: ulsprovision
  hosts: manual
  become: yes

  tasks:

# Adjust Firewall settings to allow data through.
  - name: Update firewalld services
    firewalld:
      permanent: yes
      immediate: yes
      service: "{{ item.service }}"
      state: "{{ item.state }}"
      zone: "{{ item.zone }}"
    loop:
       - { service: "http", state: "enabled", zone: "public" }
       - { service: "https", state: "enabled", zone: "public" }

  - name: Update firewalld ports
    firewalld:
      permanent: yes
      immediate: yes
      port: "{{ item.port }}/{{ item.proto }}"
      state: "{{ item.state }}"
      zone: "{{ item.zone }}"
    loop:
      - { port: "8080", proto: "tcp", state: "enabled", zone: "public" }
      - { port: "80", proto: "tcp", state: "disabled", zone: "public" }
      - { port: "9000", proto: "tcp", state: "enabled", zone: "public" }


# Install needed pre-requisites
  - name: Install java
    yum:
      name: java-1.8.0-openjdk
      state: latest

  - name: Install java-devel
    yum:
      name: java-1.8.0-openjdk-devel
      state: latest

  - name: Install tomcat
    yum:
      name: tomcat
      state: latest

  - name: Install tomcat-native
    yum:
      name: tomcat-native
      state: latest

# Extract Solr to temporary location for next steps.
  - name: Unarchive /var/tmp/solr-4.10.4.tgz
    unarchive:
      src: /home/ansiblecontrol/playbooks/resources/solr/solr-4.10.4.tgz
      dest: /var/tmp

# Create /opt/solr directory
  - name: Create /opt/solr directory
    file:
      path: /opt/solr
      state: directory
      owner: tomcat
      group: tomcat
      mode: '0755'

  - name: Create /opt/solr/collection1/data directory
    file:
      path: /opt/solr/collection1/data
      state: directory
      owner: tomcat
      group: tomcat
      mode: '0755'

# Adjust SELinux settings to allow correct file access.
  - name: Set SELinux context for /opt/solr
    sefcontext:
      target: '/opt/solr(/.*)?'
      setype: tomcat_var_lib_t
      state: present

  - name: Run restorecon on /opt/solr
    command: restorecon -rv /opt/solr

# Populate /opt/solr/ directory.
  - name: Copy contents to /opt/solr
    copy:
      src: /var/tmp/solr-4.10.4/example/solr/
      dest: /opt/solr
      owner: tomcat
      group: tomcat
      remote_src: yes

# Update solr configuration for testing
#   Note: This could probably be done with the ansible XML module.
  - name: Update Solr config Step 1 Adjust AutoCommit timing.
    replace:
      path: /opt/solr/collection1/conf/solrconfig.xml
      after: '<autoCommit>'
      before: '</autoCommit>'
      regexp: '<maxTime>(.+)</maxTime>'
      replace: '<maxTime>2000</maxTime>'
      backup: yes

#   Note: This could probably be done with the ansible XML module.
  - name: Update Solr config Step 2 Add <maxDocs> stanza.
    lineinfile:
      path: /opt/solr/collection1/conf/solrconfig.xml
      insertafter: '<autoCommit>'
      line: '<maxDocs>200</maxDocs>'

#   Note: Update <openSearcher>false</openSearcher> to true.
  - name: Update openSearcher to true.
    replace:
      path: /opt/solr/collection1/conf/solrconfig.xml
      after: '<openSearcher>'
      before: '</openSearcher>'
      regexp: 'false'
      replace: 'true'
      backup: yes

#   Note: This could probably be done with the ansible XML module.
  - name: Update Solr config Step 3 Update queryResultWindowSize
    replace:
      path: /opt/solr/collection1/conf/solrconfig.xml
      after: '<queryResultWindowSize>'
      before: '</queryResultWindowSize>'
      regexp: '^(.+)$'
      replace: '50'

#   Note: This could probably be done with the ansible XML module.
  - name: Update Solr config Step 4 Update requestDispatcher to True
    replace:
      path: /opt/solr/collection1/conf/solrconfig.xml
      after: '<requestDispatcher handleSelect="'
      before: '" >'
      regexp: '^(.+)$'
      replace: 'true'

#   Note: This had to be done with the ansible XML module.
  - name: Update Solr config Step 5 Update lst defaults.
    xml:
      path: /opt/solr/collection1/conf/solrconfig.xml
      xpath: '/config/requestHandler[@name="/select"]/lst[@name="defaults"]'
      pretty_print: yes
      input_type: xml
      add_children:
        - '<str name="fl">*</str>'
        - '<str name="q.alt">*:*</str>'
        - '<str name="qf">dc.title^5 dc.subject^3 dc.description^3 dc.creator^3 dc.contributor^3 dc.type^1 dc.relation^1 dc.publisher^1 mods_identifier_local_ms^3 ds.WARC_FILTER^1 text_nodes_HOCR_hlt^1 mods_subject_hierarchicalGeographic_region_ms^3 mods_identifier_hdl_mt^3 dc.identifier^3 PID^0.5 catch_all_fields_mt^0.1</str>'

#   Note: This is to copy the current production copy of the schema.xml file for Solr.
  - name: Copy the current schema.xml file to /opt/solr/collectoin1/conf/schema.xml
    copy:
      src: /home/ansiblecontrol/playbooks/resources/solr/schema.xml
      dest: /opt/solr/collection1/conf/schema.xml
      owner: tomcat
      group: tomcat      
      backup: yes

#
#----------------------------------------------------------------------------------------------
#   NOTE: The following is not needed due to keeping the existing production schema.xml file.
#         Maybe remove after proof of concept works as expected.
#----------------------------------------------------------------------------------------------
#
#   Note: The following is to add 'PID' as a field in /opt/solr/collection1/conf/schema.xml.
#
#  - name: Step 1 - Add <fields>
#    lineinfile:
#      path: /opt/solr/collection1/conf/schema.xml
#      insertbefore: '<!-- Valid attributes for fields:'
#      line: '<fields>'
#
#  - name: Step 2 - Add </fields>
#    lineinfile:
#      path: /opt/solr/collection1/conf/schema.xml
#      insertbefore: '<!-- Field to use to determine and enforce document uniqueness.'
#      line: '</fields>'
#
#  - name: Step 3a - Update schema.xml to include various fieldType statements.
#    xml:
#      path: /opt/solr/collection1/conf/schema.xml
#      xpath: /schema
#      pretty_print: yes
#      input_type: xml
#      backup: yes
#      add_children:
#        - '<fieldType name="sint" class="solr.SortableIntField" sortMissingLast="true" omitNorms="true"/>'
#        - '<fieldType name="slong" class="solr.SortableLongField" sortMissingLast="true" omitNorms="true"/>'
#        - '<fieldType name="sfloat" class="solr.SortableFloatField" sortMissingLast="true" omitNorms="true"/>'
#        - '<fieldType name="sdouble" class="solr.SortableDoubleField" sortMissingLast="true" omitNorms="true"/>'
#
#  - name: Step 3b - Update schema.xml to include various field statements.
#    xml:
#      path: /opt/solr/collection1/conf/schema.xml
#      xpath: /schema/fields
#      pretty_print: yes
#      input_type: xml
#      backup: yes
#      add_children:
#        - '<field name="PID" type="string" indexed="true" stored="true" required="true" />'
#        - '<field name="catch_all_fields_mt" type="text" indexed="true" stored="false" multiValued="true"/>'
#        - '<field name="timestamp" type="date" indexed="true" stored="true" default="NOW" multiValued="false"/>'
#        - '<field name="OCR_t" type="text" indexed="true" stored="false"/>'
#        - '<field name="dc.description" type="text_fgs"  indexed="true"  stored="true" multiValued="true"/>'
#
#  - name: Step 3c - Update schema.xml to include various copyField statements.
#    xml:
#      path: /opt/solr/collection1/conf/schema.xml
#      xpath: /schema
#      pretty_print: yes
#      input_type: xml
#      backup: yes
#      add_children:
#        - ''
#
#  - name: Step 3d - Update schema.xml to include various dynamicField statements.
#    xml:
#      path: /opt/solr/collection1/conf/schema.xml
#      xpath: /schema/fields
#      pretty_print: yes
#      input_type: xml
#      backup: yes
#      add_children:
#        - '<dynamicField name="*_mi"  type="sint"    indexed="true"  stored="true" multiValued="true"/>'
#        - '<dynamicField name="*_ms"  type="string"  indexed="true"  stored="true" multiValued="true"/>'
#        - '<dynamicField name="*_ml"  type="slong"   indexed="true"  stored="true" multiValued="true"/>'
#        - '<dynamicField name="*_mt"  type="text"    indexed="true"  stored="false" multiValued="true"/>'
#        - '<dynamicField name="*_hlt" type="text"    indexed="true"  stored="true" termVectors="true" termPositions="true" termOffsets="true"/>'
#        - '<dynamicField name="*_mf"  type="sfloat"  indexed="true"  stored="true" multiValued="true"/>'
#        - '<dynamicField name="*_md"  type="sdouble" indexed="true"  stored="true" multiValued="true"/>'
#        - '<dynamicField name="*_mdt" type="date"    indexed="true"  stored="true" multiValued="true"/>' 
#        - '<dynamicField name="*_et" type="edgedText" indexed="true" stored="true" multiValued="true"/>'
#        - '<dynamicField name="*_mlt" type="mappedLowerText" indexed="true" stored="true" multiValued='false'/>'
#        - '<dynamicField name="*" type="text_fgs"  indexed="true"  stored="true" multiValued="true"/>'
#
#  - name: Step 4 - Remove <uniqueKey>id</uniqueKey> from schema.xml
#    xml:
#      path: /opt/solr/collection1/conf/schema.xml
#      xpath: /schema/uniqueKey
#      state: absent
#
#  - name: Step 5 - Update schema.xml to include <uniqueKey>PID</uniqueKey>
#    xml:
#      path: /opt/solr/collection1/conf/schema.xml
#      xpath: /schema
#      pretty_print: yes
#      input_type: xml
#      add_children:
#        - '<uniqueKey>PID</uniqueKey>'

# Setup the log4j.properties file.
  - name: Copy the log4j.properties file over to /opt/solr
    copy:
      src: /var/tmp/solr-4.10.4/example/resources/log4j.properties
      dest: /opt/solr/log4j.properties
      owner: tomcat
      group: tomcat
      remote_src: yes

# Setup the stopwords.txt file. 
  - name: Copy stopwords.txt over to /opt/solr
    copy:
      src: /home/ansiblecontrol/playbooks/resources/solr/stopwords.txt
      dest: /opt/solr/stopwords.txt
      owner: tomcat
      group: tomcat

# Setup the stoprwords.txt file for collection1.
  - name: Copy stopwords.txt over to /opt/solr/collection1/conf/stopwords.txt
    copy:
      src: /home/ansiblecontrol/playbooks/resources/solr/stopwords.txt
      dest: /opt/solr/collection1/conf/stopwords.txt
      owner: tomcat
      group: tomcat

# Setup the data-config.xml and data-import-config.xml.erb files.
  - name: Copy basic-solr-config files over
    copy:
      src: "{{ item.src }}"
      dest: "{{ item.dest }}"
      owner: tomcat
      group: tomcat
    loop:
      - { src: "/home/ansiblecontrol/playbooks/resources/solr/basic-solr-config/conf/data-config.xml", dest: "/opt/solr/collection1/conf/data-config.xml" }
      - { src: "/home/ansiblecontrol/playbooks/resources/solr/basic-solr-config/conf/data-import-config.xml.erb", dest: "/opt/solr/collection1/conf/data-import-config.xml.erb" }

# Setup Tomcat to run the /solr instance.
  - name: Copy contents to /usr/share/tomcat/webapps
    copy:
      src: /var/tmp/solr-4.10.4/dist/solr-4.10.4.war
      dest: /usr/share/tomcat/webapps/solr.war
      owner: tomcat
      group: tomcat
      remote_src: yes

  - name: Create /usr/share/tomcat/webapps/solr
    file:
      path: /usr/share/tomcat/webapps/solr
      owner: tomcat
      group: tomcat
      state: directory
      mode: '0755'
 
  - name: Extract solr.war
    command:
      chdir: /usr/share/tomcat/webapps/solr
      cmd: /bin/jar -xf ../solr.war

  - name: Copy libraries over to webapp
    copy:
      src: /var/tmp/solr-4.10.4/example/lib/ext/
      dest: /usr/share/tomcat/webapps/solr/WEB-INF/lib/
      owner: tomcat
      group: tomcat
      remote_src: yes

  - name: Copy jar files over to webapp
    copy:
      src: /var/tmp/solr-4.10.4/contrib/analysis-extras/lib/icu4j-53.1.jar
      dest: /usr/share/tomcat/webapps/solr/WEB-INF/lib/
      owner: tomcat
      group: tomcat
      remote_src: yes

  - name: Copy another jar file over to webapp
    copy:
      src: /var/tmp/solr-4.10.4/contrib/analysis-extras/lucene-libs/lucene-analyzers-icu-4.10.4.jar
      dest: /usr/share/tomcat/webapps/solr/WEB-INF/lib/
      owner: tomcat
      group: tomcat
      remote_src: yes

  - name: Add line to /etc/tomcat/tomcat.conf
    lineinfile: 
      path: /etc/tomcat/tomcat.conf
      insertafter: EOF
      # From Lyrasis & our Development system:
      #line: JAVA_OPTS="-server -Xms3500m -Xmx3500m -XX:+UseParallelOldGC -Dsolr.solr.home=/opt/solr"
      # From ISLE:
      line: JAVA_OPTS="-server -Xms3500m -Xmx3500m -XX:+UseG1GC -Dsolr.solr.home=/opt/solr -XX:+UseStringDeduplication -XX:MaxGCPauseMillis=200 -XX:InitiatingHeapOccupancyPercent=70 -Djava.net.preferIPv4Stack=true -Djava.net.preferIPv4Addresses=true -Dlog4j.configuration=file:///opt/solr/log4j.properties"

  - name: Check if /etc/sysconfig/tomcat needs updated
    lineinfile:
      state: absent
      path: /etc/sysconfig/tomcat
      regexp: "^CATALINA_OPTS="
    check_mode: true
    changed_when: false
    register: check2

  - name: Add line to /etc/sysconfig/tomcat
    lineinfile:
      path: /etc/sysconfig/tomcat
      insertafter: EOF
      line: CATALINA_OPTS="-Djava.rmi.server.hostname={{ ansible_host }} -Dcom.sun.management.jmxremote -Dcom.sun.management.jmxremote.port=9000 -Dcom.sun.management.jmxremote.rmi.port=9000 -Dcom.sun.management.jmxremote.ssl=false -Dcom.sun.management.jmxremote.authenticate=false"
    when: check2.found == 0

  - name: Start Tomcat
    systemd:
      name: tomcat
      enabled: yes
      state: restarted