Ansible Role: Nexus 3 OSS

This role installs and configures Nexus Repository Manager OSS version 3.x.

All configuration can be updated by re-running the role, except for the blobstores related settings, which are immutable in nexus.

Table of Contents

Note: TOC links will not function appropriately when viewing it from ansible galaxy site. View it on github

(Created with gh-md-toc)

• Ansible Role: Nexus 3 OSS
  • Table of Contents
  • History / Credits
  • Requirements
  • Role Variables
    • General variables
    • Download dir for nexus package
    • Nexus port, context path ans listening IP
    • Nexus OS user and group
    • Nexus instance directories
    • Nexus JVM Ram setting
    • Plugin installation
    • Onboarding Wizard
    • Admin password
    • Default anonymous access
    • Public hostname
    • API access for this role
    • Branding capabalities
    • Audit capability
    • Reverse proxy setup
    • LDAP configuration
    • Privileges
    • Roles
    • Users
    • Content selectors
    • Cleanup policies
    • Blobstores and repositories
    • Scheduled tasks
    • Backups
      • Restore procedure
      • Possible limitations
    • Special maintenance/debug variables
      • Purge nexus
      • Force groovy scripts registration
      • Change admin password after first install
      • Upgrade nexus to latest version
        • Fix upgrade failing on timeout waiting for nexus port
      • Skip provisionning tasks
      • Force recursive ownership check of blobstores directories
  • Dependencies
  • Example Playbook
  • Development, Contribution and Testing
    • Contributions
    • Testing
      • Groovy syntax
      • Molecule default-xxxx scenarii
      • Molecule selinux scenario
  • License
  • Author Information

History / Credits

This role is a fork of ansible-nexus3-oss by @savoirfairelinux after they announced end of maintenance. You can have a look at the following tickets in the original repository for explanations:

• https://github.com/savoirfairelinux/ansible-nexus3-oss/issues/36
• https://github.com/savoirfairelinux/ansible-nexus3-oss/issues/38

We would like to thank the original authors for the work done.

Requirements

• Fairly Up-to-date version of ansible. We follow ansible versions during maintenance/development and will take advantage of new features if needed (and update meta/main.yml for minimum version)
• Compatible OS. This role is tested through molecule on travis CI for CentOS 8, Ubuntu Bionic (18.04), and Debian buster. Other molecule scenarios can be played locally for CentOS 7, Ubuntu Xenial (16.04), and Debian stretch
• Rsync has to be installed on the target machine (it is not needed on the host running ansible if different)
• jmespath library needs to be installed on the host running the playbook (needed for the json_query filter). See requirements.txt
• Java 8 (mandatory)
  • Oracle announced Java 8 EOL. Sonatype is now recommending openjdk8
  • For more information see nexus3 system requirements
• Apache HTTPD (optional)
  • Used to setup a SSL reverse-proxy
  • The following modules must be enabled in your configuration: mod_ssl, mod_rewrite, mod_proxy, mod_proxy_http, mod_headers.

(see Dependencies section below for matching roles on galaxy)

Role Variables

Ansible variables, along with the default values (see default/main.yml) :

General variables

    nexus_version: ''
    nexus_timezone: 'UTC'
    nexus_download_url: "http://download.sonatype.com/nexus/3"
    # nexus_download_ssl_verify: <unset>
    # nexus_version_running: <unset>

The role will install latest nexus available version by default. You may fix the version by setting the nexus_version variable. See available versions at https://www.sonatype.com/download-oss-sonatype.

If you fix the version and change it to a different one, the role will try to upgrade your installation. Make sure to change to a later version in release history. Downgrading will fail (unless you re-install from scratch using the nexus_purge special var)

If you don't fix the version and play the role on an existing installation, the current installed version will be used (detecting target of {{ nexus_installation_dir}}/nexus-latest). If you want to upgrade nexus, you will have to pass the special var nexus_upgrade=true on the ansible-playbook command line. See Upgrade nexus to latest version

If you use an older version of nexus than the lastest, you should make sure you do not use features which are not available in the installed release (e.g. yum hosted repositories for nexus < 3.8.0, git lfs repo for nexus < 3.3.0, etc.)

nexus_timezone is a Java Timezone name and can be useful in combinationwith nexus_scheduled_tasks cron expressions below.

You may change the download site for packages by tuning nexus_download_url (e.g. closed environment, proxy/cache on your network...). In this case, the automatic detection of the latest version will most likelly fail and you will have to fix the version to download. If you still want to take advantage of automatic latest version detection, a call to <your_custom_location>/latest-unix.tar.gz must return an HTTP 302 redirect to the latest available version in your cache/proxy. If your download location uses https with a self-signed certificate (or a from a private PKI) and you are having troubles getting it validated (i.e. download errors in the role) and you fully trust the target you can set nexus_download_ssl_verify: false.

nexus_version_running is a variable used internally. As such, it should never be set directly It will exist only if nexus is currently installed on the host and will register the current version prior to running the role. It can be used later in your playbook if needed (e.g. for an upgrade notification email)

Download dir for nexus package

    nexus_download_dir: '/tmp'

Directory on target where the nexus package will be downloaded.

Important note: if you intend to run the role periodically to maintain/provision your nexus install, you should make sure the downloaded files will persists between run. On RHEL/Centos specifically, you should change this dir to a location that is not cleaned up automatically. If the package file does not persit, it will be downloaded again which might cause an unnecessary restart of nexus.

Nexus port, context path ans listening IP

    nexus_default_port: 8081
    nexus_application_host: '{{ httpd_setup_enable | ternary("127.0.0.1", "0.0.0.0") }}'
    nexus_default_context_path: '/'

Listening port/ip, and context path of the java nexus process.

• the listening IP/Interface (i.e. nexus_application_host) is by default dependant on the httpd_setup_enable setting. Nexus will listen only on localhost (127.0.0.1) if reverse proxy is enabled or all configured IP (0.0.0.0) if not. You can change this setting to your actual need (i.e. don't install proxy and still bind to 127.0.0.1 only if you install your own proxy)
• nexus_default_context_path has to keep the trailing slash when set, for ex. : nexus_default_context_path: '/nexus/'.

Nexus OS user and group

    nexus_os_group: 'nexus'
    nexus_os_user: 'nexus'

User and group used to own the nexus files and run the service, those will be created by the role if absent.

    nexus_os_user_home_dir: '/home/nexus'

Allow to change the nexus user default home directory

Nexus instance directories

    nexus_installation_dir: '/opt'
    nexus_data_dir: '/var/nexus'
    nexus_tmp_dir: "{{ (ansible_os_family == 'RedHat') | ternary('/var/nexus-tmp', '/tmp/nexus') }}"

Nexus directories.

• nexus_installation_dir contains the installed executable(s)
• nexus_data_dir contains all configuration, repositories and uploaded artifacts. Custom blobstores paths outside of nexus_data_dir can be configured, see nexus_blobstores below.
• nexus_tmp_dir contains all temporary files. Default path for redhat has been moved out of /tmp to overcome potential problems with automatic cleaning procedures. See #168.

Nexus JVM Ram setting

    nexus_min_heap_size: "1200M"
    nexus_max_heap_size: "{{ nexus_min_heap_size }}"
    nexus_max_direct_memory: "2G"

These are the defaults for Nexus. Please do not modify those values unless you have read the memory section of nexus system requirements and you understand what you are doing.

As a second warning, here is an extract from the above document:

Increasing the JVM heap memory larger than recommended values in an attempt to improve performance is not recommended. This actually can have the opposite effect, causing the operating system to thrash needlessly.

Plugin installation

nexus_plugin_urls: []

Put list of urls pointing to plugins build for your Nexus version. Only *.kar bundles can be installed this way.

Onboarding Wizard

nexus_onboarding_wizard: false

Controls whether the nexus onboarding wizard runs when the admin user logs in for the first time

Admin password

    nexus_admin_password: 'changeme'

The 'admin' account password to setup. This works only on first time install by default. Please see Change admin password after first install if you want to change it later with the role.

It is strongly advised that you do not keep your password in clear text in you playbook and use ansible-vault encryption (either inline or in a separate file loaded with include_vars for example)

Default anonymous access

    nexus_anonymous_access: false

Allow anonymous access to nexus.

Public hostname

    nexus_public_hostname: 'nexus.vm'
    nexus_public_scheme: https

The fully qualified domain name and scheme under which the nexus instance will be accessible to its clients.

API access for this role

    nexus_api_hostname: localhost
    nexus_api_scheme: http
    nexus_api_validate_certs: "{{ nexus_api_scheme == 'https' }}"
    nexus_api_context_path: "{{ nexus_default_context_path }}"
    nexus_api_port: "{{ nexus_default_port }}"

These vars control how the role connects to the nexus API for provisionning. For advance usage only. You most probably do not want to change these default settings

Branding capabalities

    nexus_branding_header: ""
    nexus_branding_footer: "Last provisionned {{ ansible_date_time.iso8601 }}"

Header and footer branding, those can contain HTML.

Audit capability

    nexus_audit_enabled: false

The Auditing capability of nexus is off by default. You can turn it on by switching this to true. Please note that the audit data is stored in nexus db, persits accross reboots and is not automatically rotated/cleared.

Reverse proxy setup

    httpd_setup_enable: false
    httpd_server_name: "{{ nexus_public_hostname }}"
    httpd_default_admin_email: "[email protected]"
    httpd_ssl_certificate_file: 'files/nexus.vm.crt'
    httpd_ssl_certificate_key_file: 'files/nexus.vm.key'
    # httpd_ssl_certificate_chain_file: "{{ httpd_ssl_certificate_file }}"
    httpd_copy_ssl_files: true

Setup an SSL Reverse-proxy. This needs httpd installed. Note : when httpd_setup_enable is set to true, nexus binds by default to 127.0.0.1:8081 thus not being directly accessible on HTTP port 8081 from an external IP. (If you want to change this, you can explicitely set nexus_application_host: 0.0.0.0)

The default hostname used is nexus_public_hostname. If you need different names for whatever reason, you can set httpd_server_name to a different value.

With httpd_copy_ssl_files: true (default), the above certs must exist in your playbook dir and will be copied to the server and configured in apache. httpd_ssl_certificate_chain_file is optional and must be left unset if you do not want to configure a chain file.

If you want to use existing certificates on the server, set httpd_copy_ssl_files: false and provide the following variables

    # These specifies to the vhost where to find on the remote server file
    # system the certificate files.
    httpd_ssl_cert_file_location: "/etc/pki/tls/certs/wildcard.vm.crt"
    httpd_ssl_cert_key_location: "/etc/pki/tls/private/wildcard.vm.key"
    # httpd_ssl_cert_chain_file_location: "{{ httpd_ssl_cert_file_location }}"

httpd_ssl_cert_chain_file_location is optional and must be left unset if you do not want to configure a chain file

    httpd_default_admin_email: "[email protected]"

Set httpd default admin email address

LDAP configuration

Ldap connections and security realm are disabled by default

    nexus_ldap_realm: false
    ldap_connections: []

LDAP connection(s) setup, each item goes as follow :

    nexus_ldap_realm: true
    ldap_connections:
      - ldap_name: 'My Company LDAP' # used as a key to update the ldap config
        ldap_protocol: 'ldaps' # ldap or ldaps
        ldap_hostname: 'ldap.mycompany.com'
        ldap_port: 636
        ldap_use_trust_store: false # Wether or not to use certs in the nexus trust store
        ldap_search_base: 'dc=mycompany,dc=net'
        ldap_auth: 'none' # or simple
        ldap_auth_username: 'username' # if auth = simple
        ldap_auth_password: 'password' # if auth = simple
        ldap_user_base_dn: 'ou=users'
        ldap_user_filter: '(cn=*)' # (optional)
        ldap_user_object_class: 'inetOrgPerson'
        ldap_user_id_attribute: 'uid'
        ldap_user_real_name_attribute: 'cn'
        ldap_user_email_attribute: 'mail'
        ldap_user_subtree: false
        ldap_map_groups_as_roles: false
        ldap_group_base_dn: 'ou=groups'
        ldap_group_object_class: 'posixGroup'
        ldap_group_id_attribute: 'cn'
        ldap_group_member_attribute: 'memberUid'
        ldap_group_member_format: '${username}'
        ldap_group_subtree: false

Example LDAP config for anonymous authentication (anonymous bind), this is also the "minimal" config :

    nexus_ldap_realm: true
    ldap_connection:
      - ldap_name: 'Simplest LDAP config'
        ldap_protocol: 'ldaps'
        ldap_hostname: 'annuaire.mycompany.com'
        ldap_search_base: 'dc=mycompany,dc=net'
        ldap_port: 636
        ldap_use_trust_store: false
        ldap_user_id_attribute: 'uid'
        ldap_user_real_name_attribute: 'cn'
        ldap_user_email_attribute: 'mail'
        ldap_user_object_class: 'inetOrgPerson'

Example LDAP config for simple authentication (using a DSA account) :

    nexus_ldap_realm: true
    ldap_connections:
      - ldap_name: 'LDAP config with DSA'
        ldap_protocol: 'ldaps'
        ldap_hostname: 'annuaire.mycompany.com'
        ldap_port: 636
        ldap_use_trust_store: false
        ldap_auth: 'simple'
        ldap_auth_username: 'cn=mynexus,ou=dsa,dc=mycompany,dc=net'
        ldap_auth_password: "{{ vault_ldap_dsa_password }}" # better keep passwords in an ansible vault
        ldap_search_base: 'dc=mycompany,dc=net'
        ldap_user_base_dn: 'ou=users'
        ldap_user_object_class: 'inetOrgPerson'
        ldap_user_id_attribute: 'uid'
        ldap_user_real_name_attribute: 'cn'
        ldap_user_email_attribute: 'mail'
        ldap_user_subtree: false

Example LDAP config for simple authentication (using a DSA account) + groups mapped as roles :

    nexus_ldap_realm: true
    ldap_connections
      - ldap_name: 'LDAP config with DSA'
        ldap_protocol: 'ldaps'
        ldap_hostname: 'annuaire.mycompany.com'
        ldap_port: 636
        ldap_use_trust_store: false
        ldap_auth: 'simple'
        ldap_auth_username: 'cn=mynexus,ou=dsa,dc=mycompany,dc=net'
        ldap_auth_password: "{{ vault_ldap_dsa_password }}" # better keep passwords in an ansible vault
        ldap_search_base: 'dc=mycompany,dc=net'
        ldap_user_base_dn: 'ou=users'
        ldap_user_object_class: 'inetOrgPerson'
        ldap_user_id_attribute: 'uid'
        ldap_user_real_name_attribute: 'cn'
        ldap_user_email_attribute: 'mail'
        ldap_map_groups_as_roles: true
        ldap_group_base_dn: 'ou=groups'
        ldap_group_object_class: 'groupOfNames'
        ldap_group_id_attribute: 'cn'
        ldap_group_member_attribute: 'member'
        ldap_group_member_format: 'uid=${username},ou=users,dc=mycompany,dc=net'
        ldap_group_subtree: false

Example LDAP config for simple authentication (using a DSA account) + groups mapped as roles dynamically :

    nexus_ldap_realm: true
    ldap_connections:
      - ldap_name: 'LDAP config with DSA'
        ldap_protocol: 'ldaps'
        ldap_hostname: 'annuaire.mycompany.com'
        ldap_port: 636
        ldap_use_trust_store: false
        ldap_auth: 'simple'
        ldap_auth_username: 'cn=mynexus,ou=dsa,dc=mycompany,dc=net'
        ldap_auth_password: "{{ vault_ldap_dsa_password }}" # better keep passwords in an ansible vault
        ldap_search_base: 'dc=mycompany,dc=net'
        ldap_user_base_dn: 'ou=users'
        ldap_user_object_class: 'inetOrgPerson'
        ldap_user_id_attribute: 'uid'
        ldap_user_real_name_attribute: 'cn'
        ldap_user_email_attribute: 'mail'
        ldap_map_groups_as_roles: true
        ldap_map_groups_as_roles_type: 'dynamic'
        ldap_user_memberof_attribute: 'memberOf'

Privileges

    nexus_privileges:
      - name: all-repos-read # used as key to update a privilege
        # type: <one of application, repository-admin, repository-content-selector, repository-view, script or wildcard>
        description: 'Read & Browse access to all repos'
        repository: '*'
        actions: # can be add, browse, create, delete, edit, read or  * (all)
          - read
          - browse
        # pattern: pattern
        # domain: domain
        # script_name: name

List of the privileges to setup. Please see documentation and GUI to check out which variables should be set depending on the type of privilege.

Those items are combined with the following default values :

    _nexus_privilege_defaults:
      type: repository-view
      format: maven2
      actions:
        - read

Roles

    nexus_roles:
      - id: Developpers # can map to a LDAP group id, also used as a key to update a role
        name: developers
        description: All developers
        privileges:
          - nx-search-read
          - all-repos-read
        roles: [] # references to other role names

List of the roles to setup.

Users

    nexus_local_users: []
      # - username: jenkins # used as key to update
      #   state: present # default value if ommited, use 'absent' to remove user
      #   first_name: Jenkins
      #   last_name: CI
      #   email: [email protected]
      #   password: "s3cr3t"
      #   roles:
      #     - developers # role ID

Local (non-LDAP) users/accounts list to create in nexus. State absent will remove the user if it exists

      nexus_ldap_users: []
      # - username: j.doe
      #   state: present
      #   roles:
      #     - "nx-admin"

Ldap users/roles mappings. State absent will remove roles from the existing user if already present. Ldap users are not removed. Trying to set roles on a non existing user will result in an error.

Content selectors

  nexus_content_selectors:
  - name: docker-login
    description: Selector for docker login privilege
    search_expression: format=="docker" and path=~"/v2/"

For more info on Content selector see documentation

To use content selector add new privilege with type: repository-content-selector and proper contentSelector

- name: docker-login-privilege
  type: repository-content-selector
  contentSelector: docker-login
  description: 'Login to Docker registry'
  repository: '*'
  actions:
  - read
  - browse

Cleanup policies

nexus_repos_cleanup_policies:
#   - name: mvn_cleanup
#     format: maven2
#     mode:
#     notes: ""
#     criteria:
#       lastBlobUpdated: 60
#       lastDownloaded: 120
#       preRelease: RELEASES
#       regexKey: "foo.*"

Cleanup policies definitions. Can be added to repo definitions with the option cleanup_policies

Blobstores and repositories

    nexus_delete_default_repos: false

Delete the repositories from the nexus install initial default configuration. This step is only executed on first-time install (when nexus_data_dir has been detected empty).

    nexus_delete_default_blobstore: false

Delete the default blobstore from the nexus install initial default configuration. This can be done only if nexus_delete_default_repos: true and all configured repositories (see below) have an explicit blob_store: custom. This step is only executed on first-time install (when nexus_data_dir has been detected empty).

    nexus_blobstores: []
    # example blobstore item :
    # - name: separate-storage
    #   type: file
    #   path: /mnt/custom/path
    # - name: s3-blobstore
    #   type: S3
    #   config:
    #     bucket: s3-blobstore
    #     accessKeyId: "{{ VAULT_ENCRYPTED_KEY_ID }}"
    #     secretAccessKey: "{{ VAULT_ENCRYPTED_ACCESS_KEY }}"

Blobstores to create. A blobstore path and a repository blobstore cannot be updated after initial creation (any update here will be ignored on re-provisionning).

Configuring blobstore on S3 is provided as a convenience and is not part of the automated tests we run on travis. Please note that storing on S3 is only recommended for instances deployed on AWS.

    nexus_repos_maven_proxy:
      - name: central
        remote_url: 'https://repo1.maven.org/maven2/'
        layout_policy: permissive
        # cleanup_policies:
        #    - mvn_cleanup
        # maximum_component_age: -1
        # maximum_metadata_age: 1440
        # negative_cache_enabled: true
        # negative_cache_ttl: 1440
      - name: jboss
        remote_url: 'https://repository.jboss.org/nexus/content/groups/public-jboss/'
        # cleanup_policies:
        #    - mvn_cleanup
        # maximum_component_age: -1
        # maximum_metadata_age: 1440
        # negative_cache_enabled: true
        # negative_cache_ttl: 1440
    # example with a login/password :
    # - name: secret-remote-repo
    #   remote_url: 'https://company.com/repo/secure/private/go/away'
    #   remote_username: 'username'
    #   remote_password: 'secret'
    #   # maximum_component_age: -1
    #   # maximum_metadata_age: 1440
    #   # negative_cache_enabled: true
    #   # negative_cache_ttl: 1440

Maven proxy repositories configuration.

    nexus_repos_maven_hosted:
      - name: private-release
        version_policy: release
        write_policy: allow_once  # one of "allow", "allow_once" or "deny"
        # cleanup_policies:
        #    - mvn_cleanup

Maven hosted repositories configuration. Negative cache config is optionnal and will default to the above values if omitted.

    nexus_repos_maven_group:
      - name: public
        member_repos:
          - central
          - jboss

Maven group repositories configuration.

All three repository types are combined with the following default values :

    _nexus_repos_maven_defaults:
      blob_store: default # Note : cannot be updated once the repo has been created
      strict_content_validation: true
      version_policy: release # release, snapshot or mixed
      layout_policy: strict # strict or permissive
      write_policy: allow_once # one of "allow", "allow_once" or "deny"
      maximum_component_age: -1  # Nexus gui default. For proxies only
      maximum_metadata_age: 1440  # Nexus gui default. For proxies only
      negative_cache_enabled: true # Nexus gui default. For proxies only
      negative_cache_ttl: 1440 # Nexus gui default. For proxies only

Maven, Pypi, Docker, Raw, Rubygems, Bower, NPM, Git-LFS, yum, apt, helm, r, p2, conda and go repository types: see defaults/main.yml for these options. For historical reasons and to keep backward compatibility, maven is configured by default

      nexus_config_maven: true
      nexus_config_pypi: false
      nexus_config_docker: false
      nexus_config_raw: false
      nexus_config_rubygems: false
      nexus_config_bower: false
      nexus_config_npm: false
      nexus_config_gitlfs: false
      nexus_config_yum: false
      nexus_config_apt: false
      nexus_config_helm: false
      nexus_config_r: false
      nexus_config_p2: false
      nexus_config_conda: false
      nexus_config_go: false

These are all false unless you override them from playbook / group_var / cli, these all utilize the same mechanism as maven.

Note that you might need to enable certain security realms if you want to use other repository types than maven. These are false by default

nexus_nuget_api_key_realm: false
nexus_npm_bearer_token_realm: false
nexus_docker_bearer_token_realm: false  # required for docker anonymous access

The Remote User Realm can also be enabled with

nexus_rut_auth_realm: true

and the header can be configured by defining

nexus_rut_auth_header: "CUSTOM_HEADER"

Scheduled tasks

These are quick examples and instruction to setup scheduled tasks. For in depth information on available tasks types and schedule types, please refer to the specific section in the repo wiki

    nexus_scheduled_tasks: []
    #  #  Example task to compact blobstore :
    #  - name: compact-docker-blobstore
    #    cron: '0 0 22 * * ?'
    #    typeId: blobstore.compact
    #    task_alert_email: [email protected]  # optional
    #    taskProperties:
    #      blobstoreName: {{ nexus_blob_names.docker.blob }} # all task attributes are stored as strings by nexus internally
    #  #  Example task to purge maven snapshots
    #  - name: Purge-maven-snapshots
    #    cron: '0 50 23 * * ?'
    #    typeId: repository.maven.remove-snapshots
    #    task_alert_email: [email protected]  # optional
    #    taskProperties:
    #      repositoryName: "*"  # * for all repos. Change to a repository name if you only want a specific one
    #      minimumRetained: "2"
    #      snapshotRetentionDays: "2"
    #      gracePeriodInDays: "2"
    #    booleanTaskProperties:
    #      removeIfReleased: true
    #  #  Example task to purge unused docker manifest and images
    #  - name: Purge unused docker manifests and images
    #    cron: '0 55 23 * * ?'
    #    typeId: "repository.docker.gc"
    #    task_alert_email: [email protected]  # optional
    #    taskProperties:
    #      repositoryName: "*"  # * for all repos. Change to a repository name if you only want a specific one
    #  #  Example task to purge incomplete docker uploads
    #  - name: Purge incomplete docker uploads
    #    cron: '0 0 0 * * ?'
    #    typeId: "repository.docker.upload-purge"
    #    task_alert_email: [email protected]  # optional
    #    taskProperties:
    #      age: "24"

Scheduled tasks to setup. typeId and task-specific taskProperties/booleanTaskProperties can be guessed either:

• from the java type hierarchy of org.sonatype.nexus.scheduling.TaskDescriptorSupport
• by inspecting the task creation html form in your browser
• from peeking at the browser AJAX requests while manually configuring a task.

Task properties must be declared in the correct yaml block depending on their type:

• taskProperties for all string properties (i.e. repository names, blobstore names, time periods...).
• booleanTaskProperties for all boolean properties (i.e. mainly checkboxes in nexus create task GUI).

Backups

      nexus_backup_configure: false
      nexus_backup_schedule_type: cron
      nexus_backup_cron: '0 0 21 * * ?'  # See cron expressions definition in nexus create task gui
      # nexus_backup_start_date_time: "yyyy-MM-dd'T'HH:mm:ss"
      # nexus_backup_weekly_days: ['MON', 'TUE', 'WED', 'THU', 'FRI', 'SAT']
      # nexus_backup_monthly_days: {{ range(1,32) | list + [999] }}
      nexus_backup_dir: '/var/nexus-backup'
      nexus_backup_dir_create: true
      nexus_restore_log: '{{ nexus_backup_dir }}/nexus-restore.log'
      nexus_backup_rotate: false
      nexus_backup_rotate_first: false
      nexus_backup_keep_rotations: 4  # Keep 4 backup rotation by default (current + last 3)

Backup will not be configured unless you switch nexus_backup_configure: true. In this case, a script task will be configured in nexus.

The script task schedule will be set as cron by default and runs every day at 21:00. You can define whatever schedule you like by setting accordingly the variables nexus_backup_schedule_type, nexus_backup_cron, nexus_backup_start_date_time, nexus_backup_weekly_days and nexus_backup_monthly_days. To understand their usage depending on the type of schedule you choose, please see Scheduled tasks

See the groovy template for this task for details. This scheduled task is independent from the other nexus_scheduled_tasks you declare in your playbook

If you want to rotate backups, set nexus_backup_rotate: true and adjust the number of rotations you would like to keep with nexus_backup_keep_rotations (defaults to 4).

When using rotation, if you want to save extra disk space during the backup process, you can set nexus_backup_rotate_first: true. This will configure a pre-rotation rather than the default post-rotation. Please note than in this case, old backup(s) is/are removed before the current one is done and successful.

If you want to backup to a mounted directory (like s3fs), you can set the nexus_backup_dir_create to false.

Restore procedure

Run your playbook with parameter -e nexus_restore_point=<YYYY-MM-dd-HH-mm-ss> (e.g. 2017-12-17-21-00-00 for 17th of December 2017 at 21h00m00s)

Possible limitations

Blobstore copies are made directly from nexus by the script scheduled task. This has only been tested on rather small blobstores (less than 50Go) and should be used with caution and tested carefully on larger installations before moving to production. In any case, you are free to implement your own backup scenario outside of this role.

Special maintenance/debug variables

These are not present in defaults/main.yml and are meant to be used on the command line only for maintenance/debug reasons.

Purge nexus

** Warning: this will completely erase the current data. Make sure to backup previously if needed **

Use the nexus_purge variable if you need to restart from scratch and re-install a blank instance of nexus.

ansible-playbook -i your/inventory.ini your_nexus_playbook.yml -e nexus_purge=true

Force groovy scripts registration

This one is safe and will only make the playbook run longer if it wasn't needed

For performance sake, we use a little trick with several rsync to detect which maintenance groovy scripts need to be registered in Nexus. On some occasions (e.g. bad admin password, recovering a backup from a previous nexus instance with unregistered scripts...), this can lead to situation where the role will fail when attempting to run the needed groovy scripts.

The symptom: you get HTTP 404 errors when the role tries to run scripts like in the following example (use -v option for ansible playbook):

fatal: [nexus3-oss]: FAILED! => {"changed": false, "connection": "close", "content": "", "date": "Tue, 11 Sep 2018 07:57:44 GMT", "msg": "Status code was 404 and not [200, 204]: HTTP Error 404: Not Found", "redirected": false, "server": "Nexus/3.13.0-01 (OSS)", "status": 404, "url": "http://localhost:8081/service/rest/v1/script/update_admin_password/run", "x_content_type_options": "nosniff", "x_siesta_faultid": "914acef2-f644-4bd6-9a7d-ce19255ea3dd"}

In such cases, you can force the (re-)registration of the groovy scripts with the nexus_force_groovy_scripts_registration variable:

ansible-playbook -i your/inventory.ini your_playbook.yml -e nexus_force_groovy_scripts_registration=true

Change admin password after first install

    nexus_default_admin_password: 'admin123'

This should not be changed in your playbook. This var is filled with the default nexus admin password on first install and ensures we can change the admin password to nexus_admin_password.

If you want to change your admin password after first install, you can temporarily change this to your old password from the command line. After changing nexus_admin_password in your playbook, you can run:

ansible-playbook -i your/inventory.ini your_playbook.yml -e nexus_default_admin_password=oldPassword

Upgrade nexus to latest version

    nexus_upgrade: true

This variable has no effect if nexus_version is fixed in your vars

Unless you set this variable, the role will keep the current installed nexus version when running against an already provisioned host. Passing this extra var will trigger automatic latest nexus version detection and upgrade if a newer version is available.

Setting this var as part of your playbook breaks idempotence (i.e. your playbook will make changes to your system if a new version is available although no parameters have changed)

We strongly suggest to use this variable only as an extra var to ansible-playbook call

ansible-playbook -i your/inventory.ini your_playbook.yml -e nexus_upgrade=true

Fix upgrade failing on timeout waiting for nexus port

If you have a large nexus repository, you may occasionally see an error message when upgrading

RUNNING HANDLER [nexus3-oss : wait-for-nexus-port] *************
fatal: [nexushost]: FAILED! => {"changed": false, "elapsed": 300, "msg": "Timeout when waiting for 127.0.0.1:8081"}

This is most likely because the nexus upgrade process (i.e. migrating internal orientdb) is taking longer than the default 300 seconds. You can overcome this situation by setting a custom timeout in seconds to or/and a number of retries for the handler task.

ansible-playbook -i your/inventory.ini your_playbook.yml \
-e nexus_upgrade=true \
-e nexus_wait_for_port_timeout=600
-e nexus_wait_for_port_retries=2

Skip provisionning tasks

    nexus_run_provisionning: false

This var is unset by default and will default to true. Setting it to false will cause the role to skip all of the provisionning tasks and will therefore not create/update:

• ldap configurations
• content selectors
• privileges
• roles
• users (except checking/updating admin password)
• blobstores
• repositories
• tasks (backup will still be configured if enabled)

This can save time if you have lots of configured repositories/users/roles... and you want to play the role to simply check nexus is correctly installed, or restore a backup, or upgrade nexus version.

We strongly suggest to use this variable only as an extra var to ansible-playbook call

ansible-playbook -i your/inventory.ini your_playbook.yml -e nexus_run_provisionning=false

Force recursive ownership check of blobstores directories

Introduced in version 2.4.9

    nexus_blobstores_recurse_owner: true

In versions prior to 2.4.9, the task creating the blobstores directories was recursively checking the ownership of all files. This was not a problem on creation (where dir is empty) or with installations with small blobstores, but could lead to extremely long delays for large blobstores with lots of files.

Recursive checking of ownership has been turned off by default to prevent this extra delay. If for some reason you need to make sure all files in the blobstore directories are owned by the nexus user, you can force the check:

ansible-playbook -i your/inventory.ini your_playbook.yml -e nexus_blobstores_recurse_owner=true

Dependencies

The java and httpd requirements /can/ be fulfilled with the following galaxy roles :

• geerlingguy.java
• geerlingguy.apache

Feel free to use them or implement your own install scenario at your convenience.

Example Playbook

---
- name: Nexus
  hosts: nexus
  become: yes

  vars:
    nexus_timezone: 'Canada/Eastern'
    nexus_admin_password: "{{ vault_nexus_admin_password }}"
    nexus_public_hostname: 'nexus.vm'
    httpd_setup_enable: true
    httpd_ssl_certificate_file: "{{ vault_httpd_ssl_certificate_file }}"
    httpd_ssl_certificate_key_file: "{{ vault_httpd_ssl_certificate_key_file }}"
    ldap_connections:
      - ldap_name: 'Company LDAP'
        ldap_protocol: 'ldaps'
        ldap_hostname: 'ldap.company.com'
        ldap_port: 636
        ldap_search_base: 'dc=company,dc=net'
        ldap_user_base_dn: 'ou=users'
        ldap_user_object_class: 'inetOrgPerson'
        ldap_user_id_attribute: 'uid'
        ldap_user_real_name_attribute: 'cn'
        ldap_user_email_attribute: 'mail'
        ldap_group_base_dn: 'ou=groups'
        ldap_group_object_class: 'posixGroup'
        ldap_group_id_attribute: 'cn'
        ldap_group_member_attribute: 'memberUid'
        ldap_group_member_format: '${username}'
    nexus_privileges:
      - name: all-repos-read
        description: 'Read & Browse access to all repos'
        repository: '*'
        actions:
          - read
          - browse
      - name: company-project-deploy
        description: 'Deployments to company-project'
        repository: company-project
        actions:
          - add
          - edit
    nexus_roles:
      - id: Developpers # maps to the LDAP group
        name: developers
        description: All developers
        privileges:
          - nx-search-read
          - all-repos-read
          - company-project-deploy
        roles: []
    nexus_local_users:
      - username: jenkins # used as key to update
        first_name: Jenkins
        last_name: CI
        email: [email protected]
        password: "s3cr3t"
        roles:
          - Developpers # role ID here
    nexus_blobstores:
      - name: company-artifacts
        path: /var/nexus/blobs/company-artifacts
    nexus_scheduled_tasks:
      - name: compact-blobstore
        cron: '0 0 22 * * ?'
        typeId: blobstore.compact
        taskProperties:
          blobstoreName: 'company-artifacts'
    nexus_repos_maven_proxy:
      - name: central
        remote_url: 'https://repo1.maven.org/maven2/'
        layout_policy: permissive
      - name: alfresco
        remote_url: 'https://artifacts.alfresco.com/nexus/content/groups/private/'
        remote_username: 'secret-username'
        remote_password: "{{ vault_alfresco_private_password }}"
      - name: jboss
        remote_url: 'https://repository.jboss.org/nexus/content/groups/public-jboss/'
      - name: vaadin-addons
        remote_url: 'https://maven.vaadin.com/vaadin-addons/'
      - name: jaspersoft
        remote_url: 'https://jaspersoft.artifactoryonline.com/jaspersoft/jaspersoft-repo/'
        version_policy: mixed
    nexus_repos_maven_hosted:
      - name: company-project
        version_policy: mixed
        write_policy: allow
        blob_store: company-artifacts
    nexus_repos_maven_group:
      - name: public
        member_repos:
          - central
          - jboss
          - vaadin-addons
          - jaspersoft


  roles:
    
    
    - { role: geerlingguy.java, vars: See role doc for your distribution/version }
    # Debian/Ubuntu only
    # - { role: geerlingguy.apache, apache_create_vhosts: no, apache_mods_enabled: ["proxy.load", "proxy_http.load", "headers.load"], apache_remove_default_vhost: true, tags: ["geerlingguy.apache"] }
    # RedHat/CentOS only
    - { role: geerlingguy.apache, apache_create_vhosts: no, apache_remove_default_vhost: true, tags: ["geerlingguy.apache"] }
    - { role: ansible-thoteam.nexus3-oss, tags: ['ansible-thoteam.nexus3-oss'] }

Development, Contribution and Testing

Contributions

All contributions to this role are welcome, either for bugfixes, new features or documentation.

If you wish to contribute:

• Fork the repo under your own name/organisation through github interface
• Create a branch in your own repo with a meaningfull name. We suggest the following naming convention:
  • feat/<someFeature> for features
  • fix/<someBugFix> for bug fixes
  • docfix/<someDocFix> for documentation only fixes
• If starting an important feature change, open a pull request early describing what you want to do so we can discuss it if needed. This will prevent you from doing a lot of hard work on a lot of code for changes that we cannot finally merge.
• If there are build error on your pull request, have a look at the travis log and fix the relevant errors.

Moreover, if you have time to devote for code review, merge for realeases, etc... drop an email to [email protected] to get in touch.

Testing

This role includes tests and CI integration through travis. At time being, we test:

• groovy scripts syntax
• yaml syntax and coding standard (yamllint)
• ansible good practices (ansible lint)
• a set of basic deployments on 3 different linux platforms
  • Centos 8
  • Debian buster
  • Ubuntu bionic (18.04)

Other tests are available for older platforms but not played on CI for performance reasons:

• Centos 7
• Debian stretch
• Ubuntu xenial (16.04)

Groovy syntax

This role contains a set of groovy files used to provision nexus.

If you submit changes to groovy files, please run the groovy syntax check locally before pushing your changes

./tests/test_groovySyntax.sh

This will ensure you push groovy files with correct syntax limiting the number of check errors on travis.

You will need the groovy package installed locally to run this test.

Molecule default-xxxx scenarii

The role is tested on travis with molecule. You can run these tests locally. The best way to achieve this is through a python virtualenv. You can find some more details in requirements.txt.

# Note: the following path should be outside the working dir
virtualenv /path/to/some/pyenv
. /path/to/some/pyenv/bin/activate
pip install -r requirements.txt
molecule [create|converge|destroy|test] -s <scenario name>
deactivate

Please have a look at molecule documentation (a good start is molecule --help) for further usage.

The current proposed scenarii refer to the tested platforms (see molecule/ directory). If you launch a scenario ans leave the container running (i.e. using converge for a simple deploy), you can access the running instance from your browser at https://localhost:. See the molecule/<scenario>/molecule.yml file for detail. As a convenience, here is the correspondence between scenarii and configured ports:

• default-centos7 => https://localhost:8090
• default-centos8 => https://localhost:8095
• default-debian_buster => https://localhost:8091
• default-debian_stretch => https://localhost:8092
• default-ubuntu_16.04 => https://localhost:8093
• default-ubuntu_18.04 => https://localhost:8094

To speed up tests, molecule uses docker hub images with automated build.

• Git repo: https://github.com/docker-ThoTeam/molecule_apache_openjdk8
• Docker hub registry: https://hub.docker.com/repository/docker/thoteam/molecule_apache_openjdk8

Molecule selinux scenario

*** Warning: This scenario as been removed for the moment for molecule 3.0 compatibility reason. We it until we can decide if it can be re-introduced or not ***

We included a second molecule selinux scenario. This one is not run on travis but can be used locally to:

• test selinux integration (on centos).
• run test and access the running vms under VirtualBox on you local machine.

If you wish to use this scenario you will need

• VirtualBox
• Vagrant
• molecule

A typical workflow runs like this:

• molecule create -s selinux. Once this is complete, you will see two vagrant vms (centos7 and debian-stretch) in your VirtualBox console. These Vagrant box are taken from http://vagrant.thoteam.com
• molecule converge -s selinux will run the scenario test playbook against the two vms. You can pass additionnal variables to ansible on the command line to override playbook or default vars (e.g. molecule converge -s selinux -- -e nexus_backup_rotate=true). You can converge as many times as you want.
• You can now access the gui with https://localhost:9101 (centos7) or https://localhost:9102 (debian-stretch). You will need to add a security exception for the self signed ssl certificate. If you did not change it with a command line var above, the default role admin password is "changeme"
• When you're happy with your testing, you can recycle the used space with molecule destroy -s selinux

License

GNU GPLv3

Author Information

See: https://github.com/ansible-ThoTeam