diff options
Diffstat (limited to 'roles/openshift_node')
28 files changed, 473 insertions, 204 deletions
diff --git a/roles/openshift_node/defaults/main.yml b/roles/openshift_node/defaults/main.yml index 5904ca9bc..cc000496a 100644 --- a/roles/openshift_node/defaults/main.yml +++ b/roles/openshift_node/defaults/main.yml @@ -1,5 +1,8 @@ --- -os_firewall_allow: +r_openshift_node_firewall_enabled: "{{ os_firewall_enabled | default(True) }}" +r_openshift_node_use_firewalld: "{{ os_firewall_use_firewalld | default(False) }}" +r_openshift_node_os_firewall_deny: [] +r_openshift_node_os_firewall_allow: - service: Kubernetes kubelet port: 10250/tcp - service: http @@ -8,7 +11,18 @@ os_firewall_allow: port: 443/tcp - service: OpenShift OVS sdn port: 4789/udp - when: openshift.common.use_openshift_sdn | bool + cond: openshift.common.use_openshift_sdn | default(true) | bool - service: Calico BGP Port port: 179/tcp - when: openshift.common.use_calico | bool + cond: "{{ openshift.common.use_calico | bool }}" +- service: Kubernetes service NodePort TCP + port: "{{ openshift_node_port_range | default('') }}/tcp" + cond: "{{ openshift_node_port_range is defined }}" +- service: Kubernetes service NodePort UDP + port: "{{ openshift_node_port_range | default('') }}/udp" + cond: "{{ openshift_node_port_range is defined }}" + +oreg_url: '' +oreg_host: "{{ oreg_url.split('/')[0] if '.' in oreg_url.split('/')[0] else '' }}" +oreg_auth_credentials_path: "{{ openshift.common.data_dir }}/.docker" +oreg_auth_credentials_replace: False diff --git a/roles/openshift_node/handlers/main.yml b/roles/openshift_node/handlers/main.yml index a6bd12d4e..f2c45a4bd 100644 --- a/roles/openshift_node/handlers/main.yml +++ b/roles/openshift_node/handlers/main.yml @@ -3,10 +3,15 @@ systemd: name: openvswitch state: restarted - when: (not skip_node_svc_handlers | default(False) | bool) and not (ovs_service_status_changed | default(false) | bool) and openshift.common.use_openshift_sdn | bool + when: (not skip_node_svc_handlers | default(False) | bool) and not (ovs_service_status_changed | default(false) | bool) and openshift.common.use_openshift_sdn | default(true) | bool + register: l_openshift_node_stop_openvswitch_result + until: not l_openshift_node_stop_openvswitch_result | failed + retries: 3 + delay: 30 notify: - restart openvswitch pause + - name: restart openvswitch pause pause: seconds=15 when: (not skip_node_svc_handlers | default(False) | bool) and openshift.common.is_containerized | bool @@ -15,7 +20,13 @@ systemd: name: "{{ openshift.common.service_type }}-node" state: restarted - when: (not skip_node_svc_handlers | default(False) | bool) and not (node_service_status_changed | default(false) | bool) + register: l_openshift_node_restart_node_result + until: not l_openshift_node_restart_node_result | failed + retries: 3 + delay: 30 + when: + - (not skip_node_svc_handlers | default(False) | bool) + - not (node_service_status_changed | default(false) | bool) - name: reload sysctl.conf command: /sbin/sysctl -p diff --git a/roles/openshift_node/meta/main.yml b/roles/openshift_node/meta/main.yml index 3b7e8126a..06373de04 100644 --- a/roles/openshift_node/meta/main.yml +++ b/roles/openshift_node/meta/main.yml @@ -12,37 +12,13 @@ galaxy_info: categories: - cloud dependencies: +- role: openshift_node_facts - role: lib_openshift +- role: lib_os_firewall - role: openshift_common - role: openshift_clock - role: openshift_docker - role: openshift_node_certificates - role: openshift_cloud_provider -- role: os_firewall - os_firewall_allow: - - service: Kubernetes kubelet - port: 10250/tcp - - service: http - port: 80/tcp - - service: https - port: 443/tcp -- role: os_firewall - os_firewall_allow: - - service: OpenShift OVS sdn - port: 4789/udp - when: openshift.common.use_openshift_sdn | bool -- role: os_firewall - os_firewall_allow: - - service: Calico BGP Port - port: 179/tcp - when: openshift.common.use_calico | bool - -- role: os_firewall - os_firewall_allow: - - service: Kubernetes service NodePort TCP - port: "{{ openshift_node_port_range | default('') }}/tcp" - - service: Kubernetes service NodePort UDP - port: "{{ openshift_node_port_range | default('') }}/udp" - when: openshift_node_port_range is defined - role: openshift_node_dnsmasq when: openshift.common.use_dnsmasq | bool diff --git a/roles/openshift_node/tasks/config/configure-node-settings.yml b/roles/openshift_node/tasks/config/configure-node-settings.yml new file mode 100644 index 000000000..1186062eb --- /dev/null +++ b/roles/openshift_node/tasks/config/configure-node-settings.yml @@ -0,0 +1,16 @@ +--- +- name: Configure Node settings + lineinfile: + dest: /etc/sysconfig/{{ openshift.common.service_type }}-node + regexp: "{{ item.regex }}" + line: "{{ item.line }}" + create: true + with_items: + - regex: '^OPTIONS=' + line: "OPTIONS=--loglevel={{ openshift.node.debug_level | default(2) }}" + - regex: '^CONFIG_FILE=' + line: "CONFIG_FILE={{ openshift.common.config_base }}/node/node-config.yaml" + - regex: '^IMAGE_VERSION=' + line: "IMAGE_VERSION={{ openshift_image_tag }}" + notify: + - restart node diff --git a/roles/openshift_node/tasks/config/configure-proxy-settings.yml b/roles/openshift_node/tasks/config/configure-proxy-settings.yml new file mode 100644 index 000000000..d60794305 --- /dev/null +++ b/roles/openshift_node/tasks/config/configure-proxy-settings.yml @@ -0,0 +1,17 @@ +--- +- name: Configure Proxy Settings + lineinfile: + dest: /etc/sysconfig/{{ openshift.common.service_type }}-node + regexp: "{{ item.regex }}" + line: "{{ item.line }}" + create: true + with_items: + - regex: '^HTTP_PROXY=' + line: "HTTP_PROXY={{ openshift.common.http_proxy | default('') }}" + - regex: '^HTTPS_PROXY=' + line: "HTTPS_PROXY={{ openshift.common.https_proxy | default('') }}" + - regex: '^NO_PROXY=' + line: "NO_PROXY={{ openshift.common.no_proxy | default([]) }},{{ openshift.common.portal_net }},{{ hostvars[groups.oo_first_master.0].openshift.master.sdn_cluster_network_cidr }}" + when: ('http_proxy' in openshift.common and openshift.common.http_proxy != '') + notify: + - restart node diff --git a/roles/openshift_node/tasks/config/install-node-deps-docker-service-file.yml b/roles/openshift_node/tasks/config/install-node-deps-docker-service-file.yml new file mode 100644 index 000000000..ee91a88ab --- /dev/null +++ b/roles/openshift_node/tasks/config/install-node-deps-docker-service-file.yml @@ -0,0 +1,8 @@ +--- +- name: Install Node dependencies docker service file + template: + dest: "/etc/systemd/system/{{ openshift.common.service_type }}-node-dep.service" + src: openshift.docker.node.dep.service + notify: + - reload systemd units + - restart node diff --git a/roles/openshift_node/tasks/config/install-node-docker-service-file.yml b/roles/openshift_node/tasks/config/install-node-docker-service-file.yml new file mode 100644 index 000000000..f92ff79b5 --- /dev/null +++ b/roles/openshift_node/tasks/config/install-node-docker-service-file.yml @@ -0,0 +1,8 @@ +--- +- name: Install Node docker service file + template: + dest: "/etc/systemd/system/{{ openshift.common.service_type }}-node.service" + src: openshift.docker.node.service + notify: + - reload systemd units + - restart node diff --git a/roles/openshift_node/tasks/config/install-ovs-docker-service-file.yml b/roles/openshift_node/tasks/config/install-ovs-docker-service-file.yml new file mode 100644 index 000000000..c2c5ea1d4 --- /dev/null +++ b/roles/openshift_node/tasks/config/install-ovs-docker-service-file.yml @@ -0,0 +1,8 @@ +--- +- name: Install OpenvSwitch docker service file + template: + dest: "/etc/systemd/system/openvswitch.service" + src: openvswitch.docker.service + notify: + - reload systemd units + - restart openvswitch diff --git a/roles/openshift_node/tasks/config/install-ovs-service-env-file.yml b/roles/openshift_node/tasks/config/install-ovs-service-env-file.yml new file mode 100644 index 000000000..1d75a3355 --- /dev/null +++ b/roles/openshift_node/tasks/config/install-ovs-service-env-file.yml @@ -0,0 +1,8 @@ +--- +- name: Create the openvswitch service env file + template: + src: openvswitch.sysconfig.j2 + dest: /etc/sysconfig/openvswitch + notify: + - reload systemd units + - restart openvswitch diff --git a/roles/openshift_node/tasks/config/workaround-bz1331590-ovs-oom-fix.yml b/roles/openshift_node/tasks/config/workaround-bz1331590-ovs-oom-fix.yml new file mode 100644 index 000000000..5df1abc79 --- /dev/null +++ b/roles/openshift_node/tasks/config/workaround-bz1331590-ovs-oom-fix.yml @@ -0,0 +1,13 @@ +--- +# May be a temporary workaround. +# https://bugzilla.redhat.com/show_bug.cgi?id=1331590 +- name: Create OpenvSwitch service.d directory + file: path=/etc/systemd/system/openvswitch.service.d/ state=directory + +- name: Install OpenvSwitch service OOM fix + template: + dest: "/etc/systemd/system/openvswitch.service.d/01-avoid-oom.conf" + src: openvswitch-avoid-oom.conf + notify: + - reload systemd units + - restart openvswitch diff --git a/roles/openshift_node/tasks/firewall.yml b/roles/openshift_node/tasks/firewall.yml new file mode 100644 index 000000000..255aa886a --- /dev/null +++ b/roles/openshift_node/tasks/firewall.yml @@ -0,0 +1,40 @@ +--- +- when: r_openshift_node_firewall_enabled | bool and not r_openshift_node_use_firewalld | bool + block: + - name: Add iptables allow rules + os_firewall_manage_iptables: + name: "{{ item.service }}" + action: add + protocol: "{{ item.port.split('/')[1] }}" + port: "{{ item.port.split('/')[0] }}" + when: item.cond | default(True) + with_items: "{{ r_openshift_node_os_firewall_allow }}" + + - name: Remove iptables rules + os_firewall_manage_iptables: + name: "{{ item.service }}" + action: remove + protocol: "{{ item.port.split('/')[1] }}" + port: "{{ item.port.split('/')[0] }}" + when: item.cond | default(True) + with_items: "{{ r_openshift_node_os_firewall_deny }}" + +- when: r_openshift_node_firewall_enabled | bool and r_openshift_node_use_firewalld | bool + block: + - name: Add firewalld allow rules + firewalld: + port: "{{ item.port }}" + permanent: true + immediate: true + state: enabled + when: item.cond | default(True) + with_items: "{{ r_openshift_node_os_firewall_allow }}" + + - name: Remove firewalld allow rules + firewalld: + port: "{{ item.port }}" + permanent: true + immediate: true + state: disabled + when: item.cond | default(True) + with_items: "{{ r_openshift_node_os_firewall_deny }}" diff --git a/roles/openshift_node/tasks/main.yml b/roles/openshift_node/tasks/main.yml index 573051504..525dd1d1a 100644 --- a/roles/openshift_node/tasks/main.yml +++ b/roles/openshift_node/tasks/main.yml @@ -2,9 +2,13 @@ # TODO: allow for overriding default ports where possible - fail: msg: "SELinux is disabled, This deployment type requires that SELinux is enabled." - when: > - (not ansible_selinux or ansible_selinux.status != 'enabled') and - deployment_type in ['enterprise', 'online', 'atomic-enterprise', 'openshift-enterprise'] + when: + - (not ansible_selinux or ansible_selinux.status != 'enabled') and deployment_type in ['enterprise', 'online', 'atomic-enterprise', 'openshift-enterprise'] + - not openshift_use_crio | default(false) + +- name: setup firewall + include: firewall.yml + static: yes - name: Set node facts openshift_facts: @@ -66,31 +70,30 @@ - openshift_disable_swap | default(true) | bool # End Disable Swap Block -# We have to add tuned-profiles in the same transaction otherwise we run into depsolving -# problems because the rpms don't pin the version properly. This was fixed in 3.1 packaging. - name: Install Node package package: - name: "{{ openshift.common.service_type }}-node{{ openshift_pkg_version | default('') | oo_image_tag_to_rpm_version(include_dash=True) }},tuned-profiles-{{ openshift.common.service_type }}-node{{ openshift_pkg_version | default('') | oo_image_tag_to_rpm_version(include_dash=True) }}" + name: "{{ openshift.common.service_type }}-node{{ openshift_pkg_version | default('') | oo_image_tag_to_rpm_version(include_dash=True) }}" state: present when: not openshift.common.is_containerized | bool -- name: Check for tuned package - command: rpm -q tuned - args: - warn: no - register: tuned_installed - changed_when: false - failed_when: false - -- name: Set atomic-guest tuned profile - command: "tuned-adm profile atomic-guest" - when: tuned_installed.rc == 0 and openshift.common.is_atomic | bool +- name: setup tuned + include: tuned.yml + static: yes - name: Install sdn-ovs package package: name: "{{ openshift.common.service_type }}-sdn-ovs{{ openshift_pkg_version | oo_image_tag_to_rpm_version(include_dash=True) }}" state: present - when: openshift.common.use_openshift_sdn and not openshift.common.is_containerized | bool + when: + - openshift.common.use_openshift_sdn | default(true) | bool + - not openshift.common.is_containerized | bool + +- name: Restart cri-o + systemd: + name: cri-o + enabled: yes + state: restarted + when: openshift_use_crio | default(false) - name: Install conntrack-tools package package: @@ -118,8 +121,14 @@ name: openvswitch.service enabled: yes state: started - when: openshift.common.is_containerized | bool and openshift.common.use_openshift_sdn | bool + daemon_reload: yes + when: + - openshift.common.is_containerized | bool + - openshift.common.use_openshift_sdn | default(true) | bool register: ovs_start_result + until: not ovs_start_result | failed + retries: 3 + delay: 30 - set_fact: ovs_service_status_changed: "{{ ovs_start_result | changed }}" @@ -141,6 +150,21 @@ notify: - restart node +- name: Check for credentials file for registry auth + stat: + path: "{{oreg_auth_credentials_path }}" + when: + - oreg_auth_user is defined + register: node_oreg_auth_credentials_stat + +- name: Create credentials for registry auth + command: "docker --config={{ oreg_auth_credentials_path }} login -u {{ oreg_auth_user }} -p {{ oreg_auth_password }} {{ oreg_host }}" + when: + - oreg_auth_user is defined + - (not node_oreg_auth_credentials_stat.stat.exists or oreg_auth_credentials_replace) | bool + notify: + - restart node + - name: Configure AWS Cloud Provider Settings lineinfile: dest: /etc/sysconfig/{{ openshift.common.service_type }}-node @@ -212,15 +236,27 @@ state: started when: openshift.common.is_containerized | bool + - name: Start and enable node systemd: name: "{{ openshift.common.service_type }}-node" enabled: yes state: started + daemon_reload: yes register: node_start_result until: not node_start_result | failed retries: 1 delay: 30 + ignore_errors: true + +- name: Dump logs from node service if it failed + command: journalctl --no-pager -n 100 -u {{ openshift.common.service_type }}-node + when: node_start_result | failed + +- name: Abort if node failed to start + fail: + msg: Node failed to start please inspect the logs and try again + when: node_start_result | failed - set_fact: node_service_status_changed: "{{ node_start_result | changed }}" diff --git a/roles/openshift_node/tasks/node_system_container.yml b/roles/openshift_node/tasks/node_system_container.yml index d99f657bc..b2dceedbe 100644 --- a/roles/openshift_node/tasks/node_system_container.yml +++ b/roles/openshift_node/tasks/node_system_container.yml @@ -1,12 +1,12 @@ --- - name: Pre-pull node system container image command: > - atomic pull --storage=ostree {{ openshift.common.system_images_registry }}/{{ openshift.node.node_system_image }}:{{ openshift_image_tag }} + atomic pull --storage=ostree {{ 'docker:' if openshift.common.system_images_registry == 'docker' else openshift.common.system_images_registry + '/' }}{{ openshift.node.node_system_image }}:{{ openshift_image_tag }} register: pull_result changed_when: "'Pulling layer' in pull_result.stdout" - name: Install or Update node system container oc_atomic_container: name: "{{ openshift.common.service_type }}-node" - image: "{{ openshift.common.system_images_registry }}/{{ openshift.node.node_system_image }}:{{ openshift_image_tag }}" + image: "{{ 'docker:' if openshift.common.system_images_registry == 'docker' else openshift.common.system_images_registry + '/' }}{{ openshift.node.node_system_image }}:{{ openshift_image_tag }}" state: latest diff --git a/roles/openshift_node/tasks/openvswitch_system_container.yml b/roles/openshift_node/tasks/openvswitch_system_container.yml index 8cfa5a026..e09063aa5 100644 --- a/roles/openshift_node/tasks/openvswitch_system_container.yml +++ b/roles/openshift_node/tasks/openvswitch_system_container.yml @@ -1,12 +1,25 @@ --- +- set_fact: + l_use_crio: "{{ openshift_use_crio | default(false) }}" + +- set_fact: + l_service_name: "cri-o" + when: l_use_crio + +- set_fact: + l_service_name: "{{ openshift.docker.service_name }}" + when: not l_use_crio + - name: Pre-pull OpenVSwitch system container image command: > - atomic pull --storage=ostree {{ openshift.common.system_images_registry }}/{{ openshift.node.ovs_system_image }}:{{ openshift_image_tag }} + atomic pull --storage=ostree {{ 'docker:' if openshift.common.system_images_registry == 'docker' else openshift.common.system_images_registry + '/' }}{{ openshift.node.ovs_system_image }}:{{ openshift_image_tag }} register: pull_result changed_when: "'Pulling layer' in pull_result.stdout" - name: Install or Update OpenVSwitch system container oc_atomic_container: name: openvswitch - image: "{{ openshift.common.system_images_registry }}/{{ openshift.node.ovs_system_image }}:{{ openshift_image_tag }}" + image: "{{ 'docker:' if openshift.common.system_images_registry == 'docker' else openshift.common.system_images_registry + '/' }}{{ openshift.node.ovs_system_image }}:{{ openshift_image_tag }}" state: latest + values: + - "DOCKER_SERVICE={{ l_service_name }}" diff --git a/roles/openshift_node/tasks/storage_plugins/glusterfs.yml b/roles/openshift_node/tasks/storage_plugins/glusterfs.yml index 7d8c42ee2..1b8a7ad50 100644 --- a/roles/openshift_node/tasks/storage_plugins/glusterfs.yml +++ b/roles/openshift_node/tasks/storage_plugins/glusterfs.yml @@ -3,30 +3,52 @@ package: name=glusterfs-fuse state=present when: not openshift.common.is_atomic | bool -- name: Check for existence of virt_use_fusefs seboolean - command: getsebool virt_use_fusefs - register: virt_use_fusefs_output - when: ansible_selinux and ansible_selinux.status == "enabled" +- name: Check for existence of fusefs sebooleans + command: getsebool {{ item }} + register: fusefs_getsebool_status + when: + - ansible_selinux + - ansible_selinux.status == "enabled" failed_when: false changed_when: false + with_items: + - virt_use_fusefs + - virt_sandbox_use_fusefs - name: Set seboolean to allow gluster storage plugin access from containers seboolean: - name: virt_use_fusefs + name: "{{ item.item }}" state: yes persistent: yes - when: ansible_selinux and ansible_selinux.status == "enabled" and virt_use_fusefs_output.rc == 0 + when: + - ansible_selinux + - ansible_selinux.status == "enabled" + - item.rc == 0 + # We need to detect whether or not the boolean is an alias, since `seboolean` + # will error if it is an alias. We do this by inspecting stdout for the boolean name, + # since getsebool prints the resolved name. (At some point Ansible's seboolean module + # should learn to deal with aliases) + - item.item in item.stdout # Boolean does not have an alias. + - ansible_python_version | version_compare('3', '<') + with_items: "{{ fusefs_getsebool_status.results }}" -- name: Check for existence of virt_sandbox_use_fusefs seboolean - command: getsebool virt_sandbox_use_fusefs - register: virt_sandbox_use_fusefs_output - when: ansible_selinux and ansible_selinux.status == "enabled" - failed_when: false - changed_when: false - -- name: Set seboolean to allow gluster storage plugin access from containers(sandbox) - seboolean: - name: virt_sandbox_use_fusefs - state: yes - persistent: yes - when: ansible_selinux and ansible_selinux.status == "enabled" and virt_sandbox_use_fusefs_output.rc == 0 +# Workaround for https://github.com/openshift/openshift-ansible/issues/4438 +# Use command module rather than seboolean module to set GlusterFS booleans. +# TODO: Remove this task and the ansible_python_version comparison in +# the previous task when the problem has been addressed in current +# ansible release. +- name: Set seboolean to allow gluster storage plugin access from containers (python 3) + command: > + setsebool -P {{ item.item }} on + when: + - ansible_selinux + - ansible_selinux.status == "enabled" + - item.rc == 0 + # We need to detect whether or not the boolean is an alias, since `seboolean` + # will error if it is an alias. We do this by inspecting stdout for the boolean name, + # since getsebool prints the resolved name. (At some point Ansible's seboolean module + # should learn to deal with aliases) + - item.item in item.stdout # Boolean does not have an alias. + - ('--> off' in item.stdout) # Boolean is currently off. + - ansible_python_version | version_compare('3', '>=') + with_items: "{{ fusefs_getsebool_status.results }}" diff --git a/roles/openshift_node/tasks/storage_plugins/nfs.yml b/roles/openshift_node/tasks/storage_plugins/nfs.yml index d40ae66cb..7e1035893 100644 --- a/roles/openshift_node/tasks/storage_plugins/nfs.yml +++ b/roles/openshift_node/tasks/storage_plugins/nfs.yml @@ -3,24 +3,52 @@ package: name=nfs-utils state=present when: not openshift.common.is_atomic | bool -- name: Check for existence of seboolean +- name: Check for existence of nfs sebooleans command: getsebool {{ item }} - register: getsebool_status - when: ansible_selinux and ansible_selinux.status == "enabled" + register: nfs_getsebool_status + when: + - ansible_selinux + - ansible_selinux.status == "enabled" failed_when: false changed_when: false with_items: - - virt_use_nfs - - virt_sandbox_use_nfs + - virt_use_nfs + - virt_sandbox_use_nfs - name: Set seboolean to allow nfs storage plugin access from containers seboolean: name: "{{ item.item }}" state: yes persistent: yes + when: + - ansible_selinux + - ansible_selinux.status == "enabled" + - item.rc == 0 # We need to detect whether or not the boolean is an alias, since `seboolean` # will error if it is an alias. We do this by inspecting stdout for the boolean name, # since getsebool prints the resolved name. (At some point Ansible's seboolean module # should learn to deal with aliases) - when: ansible_selinux and ansible_selinux.status == "enabled" and item.rc == 0 and item.stdout.find(item.item) != -1 - with_items: "{{ getsebool_status.results }}" + - item.item in item.stdout # Boolean does not have an alias. + - ansible_python_version | version_compare('3', '<') + with_items: "{{ nfs_getsebool_status.results }}" + +# Workaround for https://github.com/openshift/openshift-ansible/issues/4438 +# Use command module rather than seboolean module to set NFS booleans. +# TODO: Remove this task and the ansible_python_version comparison in +# the previous task when the problem has been addressed in current +# ansible release. +- name: Set seboolean to allow nfs storage plugin access from containers (python 3) + command: > + setsebool -P {{ item.item }} on + when: + - ansible_selinux + - ansible_selinux.status == "enabled" + - item.rc == 0 + # We need to detect whether or not the boolean is an alias, since `seboolean` + # will error if it is an alias. We do this by inspecting stdout for the boolean name, + # since getsebool prints the resolved name. (At some point Ansible's seboolean module + # should learn to deal with aliases) + - item.item in item.stdout # Boolean does not have an alias. + - ('--> off' in item.stdout) # Boolean is currently off. + - ansible_python_version | version_compare('3', '>=') + with_items: "{{ nfs_getsebool_status.results }}" diff --git a/roles/openshift_node/tasks/systemd_units.yml b/roles/openshift_node/tasks/systemd_units.yml index e3ce5df3d..b86bb1549 100644 --- a/roles/openshift_node/tasks/systemd_units.yml +++ b/roles/openshift_node/tasks/systemd_units.yml @@ -2,15 +2,8 @@ # This file is included both in the openshift_master role and in the upgrade # playbooks. -- name: Install Node dependencies docker service file - template: - dest: "/etc/systemd/system/{{ openshift.common.service_type }}-node-dep.service" - src: openshift.docker.node.dep.service - register: install_node_dep_result +- include: config/install-node-deps-docker-service-file.yml when: openshift.common.is_containerized | bool - notify: - - reload systemd units - - restart node - block: - name: Pre-pull node image @@ -19,14 +12,7 @@ register: pull_result changed_when: "'Downloaded newer image' in pull_result.stdout" - - name: Install Node docker service file - template: - dest: "/etc/systemd/system/{{ openshift.common.service_type }}-node.service" - src: openshift.docker.node.service - register: install_node_result - notify: - - reload systemd units - - restart node + - include: config/install-node-docker-service-file.yml when: - openshift.common.is_containerized | bool - not openshift.common.is_node_system_container | bool @@ -34,22 +20,14 @@ - name: Install Node service file template: dest: "/etc/systemd/system/{{ openshift.common.service_type }}-node.service" - src: "{{ openshift.common.service_type }}-node.service.j2" - register: install_node_result + src: "node.service.j2" when: not openshift.common.is_containerized | bool notify: - reload systemd units - restart node -- name: Create the openvswitch service env file - template: - src: openvswitch.sysconfig.j2 - dest: /etc/sysconfig/openvswitch +- include: config/install-ovs-service-env-file.yml when: openshift.common.is_containerized | bool - register: install_ovs_sysconfig - notify: - - reload systemd units - - restart openvswitch - name: Install Node system container include: node_system_container.yml @@ -64,22 +42,9 @@ - openshift.common.is_containerized | bool - openshift.common.is_openvswitch_system_container | bool -# May be a temporary workaround. -# https://bugzilla.redhat.com/show_bug.cgi?id=1331590 -- name: Create OpenvSwitch service.d directory - file: path=/etc/systemd/system/openvswitch.service.d/ state=directory +- include: config/workaround-bz1331590-ovs-oom-fix.yml when: openshift.common.use_openshift_sdn | default(true) | bool -- name: Install OpenvSwitch service OOM fix - template: - dest: "/etc/systemd/system/openvswitch.service.d/01-avoid-oom.conf" - src: openvswitch-avoid-oom.conf - when: openshift.common.use_openshift_sdn | default(true) | bool - register: install_oom_fix_result - notify: - - reload systemd units - - restart openvswitch - - block: - name: Pre-pull openvswitch image command: > @@ -87,47 +52,11 @@ register: pull_result changed_when: "'Downloaded newer image' in pull_result.stdout" - - name: Install OpenvSwitch docker service file - template: - dest: "/etc/systemd/system/openvswitch.service" - src: openvswitch.docker.service - notify: - - reload systemd units - - restart openvswitch + - include: config/install-ovs-docker-service-file.yml when: - openshift.common.is_containerized | bool - openshift.common.use_openshift_sdn | default(true) | bool - not openshift.common.is_openvswitch_system_container | bool -- name: Configure Node settings - lineinfile: - dest: /etc/sysconfig/{{ openshift.common.service_type }}-node - regexp: "{{ item.regex }}" - line: "{{ item.line }}" - create: true - with_items: - - regex: '^OPTIONS=' - line: "OPTIONS=--loglevel={{ openshift.node.debug_level | default(2) }}" - - regex: '^CONFIG_FILE=' - line: "CONFIG_FILE={{ openshift.common.config_base }}/node/node-config.yaml" - - regex: '^IMAGE_VERSION=' - line: "IMAGE_VERSION={{ openshift_image_tag }}" - notify: - - restart node - -- name: Configure Proxy Settings - lineinfile: - dest: /etc/sysconfig/{{ openshift.common.service_type }}-node - regexp: "{{ item.regex }}" - line: "{{ item.line }}" - create: true - with_items: - - regex: '^HTTP_PROXY=' - line: "HTTP_PROXY={{ openshift.common.http_proxy | default('') }}" - - regex: '^HTTPS_PROXY=' - line: "HTTPS_PROXY={{ openshift.common.https_proxy | default('') }}" - - regex: '^NO_PROXY=' - line: "NO_PROXY={{ openshift.common.no_proxy | default([]) }},{{ openshift.common.portal_net }},{{ hostvars[groups.oo_first_master.0].openshift.master.sdn_cluster_network_cidr }}" - when: ('http_proxy' in openshift.common and openshift.common.http_proxy != '') - notify: - - restart node +- include: config/configure-node-settings.yml +- include: config/configure-proxy-settings.yml diff --git a/roles/openshift_node/tasks/tuned.yml b/roles/openshift_node/tasks/tuned.yml new file mode 100644 index 000000000..425bf6a26 --- /dev/null +++ b/roles/openshift_node/tasks/tuned.yml @@ -0,0 +1,41 @@ +--- +- name: Check for tuned package + command: rpm -q tuned + args: + warn: no + register: tuned_installed + changed_when: false + failed_when: false + +- name: Tuned service setup + block: + - name: Set tuned OpenShift variables + set_fact: + openshift_tuned_guest_profile: "{{ 'atomic-guest' if openshift.common.is_atomic else 'virtual-guest' }}" + tuned_etc_directory: '/etc/tuned' + tuned_templates_source: '../templates/tuned' + + - name: Ensure directory structure exists + file: + state: directory + dest: '{{ tuned_etc_directory }}/{{ item.path }}' + with_filetree: '{{ tuned_templates_source }}' + when: item.state == 'directory' + + - name: Ensure files are populated from templates + template: + src: '{{ item.src }}' + dest: '{{ tuned_etc_directory }}/{{ item.path }}' + with_filetree: '{{ tuned_templates_source }}' + when: item.state == 'file' + + - name: Make tuned use the recommended tuned profile on restart + file: path=/etc/tuned/active_profile state=absent + + - name: Restart tuned service + systemd: + state: restarted + daemon_reload: yes + name: tuned + + when: tuned_installed.rc == 0 | bool diff --git a/roles/openshift_node/templates/atomic-openshift-node.service.j2 b/roles/openshift_node/templates/atomic-openshift-node.service.j2 deleted file mode 100644 index 80232094a..000000000 --- a/roles/openshift_node/templates/atomic-openshift-node.service.j2 +++ /dev/null @@ -1,22 +0,0 @@ -[Unit] -Description=Atomic OpenShift Node -After={{ openshift.docker.service_name }}.service -After=openvswitch.service -Wants={{ openshift.docker.service_name }}.service -Documentation=https://github.com/openshift/origin - -[Service] -Type=notify -EnvironmentFile=/etc/sysconfig/atomic-openshift-node -Environment=GOTRACEBACK=crash -ExecStart=/usr/bin/openshift start node --config=${CONFIG_FILE} $OPTIONS -LimitNOFILE=65536 -LimitCORE=infinity -WorkingDirectory=/var/lib/origin/ -SyslogIdentifier=atomic-openshift-node -Restart=always -RestartSec=5s -OOMScoreAdjust=-999 - -[Install] -WantedBy=multi-user.target diff --git a/roles/openshift_node/templates/node.service.j2 b/roles/openshift_node/templates/node.service.j2 new file mode 100644 index 000000000..0856737f6 --- /dev/null +++ b/roles/openshift_node/templates/node.service.j2 @@ -0,0 +1,32 @@ +[Unit] +Description=OpenShift Node +After={{ openshift.docker.service_name }}.service +Wants=openvswitch.service +After=ovsdb-server.service +After=ovs-vswitchd.service +Wants={{ openshift.docker.service_name }}.service +Documentation=https://github.com/openshift/origin +Requires=dnsmasq.service +After=dnsmasq.service +{% if openshift_use_crio|default(false) %}Wants=cri-o.service{% endif %} + +[Service] +Type=notify +EnvironmentFile=/etc/sysconfig/{{ openshift.common.service_type }}-node +Environment=GOTRACEBACK=crash +ExecStartPre=/usr/bin/cp /etc/origin/node/node-dnsmasq.conf /etc/dnsmasq.d/ +ExecStartPre=/usr/bin/dbus-send --system --dest=uk.org.thekelleys.dnsmasq /uk/org/thekelleys/dnsmasq uk.org.thekelleys.SetDomainServers array:string:/in-addr.arpa/127.0.0.1,/{{ openshift.common.dns_domain }}/127.0.0.1 +ExecStopPost=/usr/bin/rm /etc/dnsmasq.d/node-dnsmasq.conf +ExecStopPost=/usr/bin/dbus-send --system --dest=uk.org.thekelleys.dnsmasq /uk/org/thekelleys/dnsmasq uk.org.thekelleys.SetDomainServers array:string: +ExecStart=/usr/bin/openshift start node --config=${CONFIG_FILE} $OPTIONS +LimitNOFILE=65536 +LimitCORE=infinity +WorkingDirectory=/var/lib/origin/ +SyslogIdentifier={{ openshift.common.service_type }}-node +Restart=always +RestartSec=5s +TimeoutStartSec=300 +OOMScoreAdjust=-999 + +[Install] +WantedBy=multi-user.target diff --git a/roles/openshift_node/templates/node.yaml.v1.j2 b/roles/openshift_node/templates/node.yaml.v1.j2 index f2f929232..711afcadb 100644 --- a/roles/openshift_node/templates/node.yaml.v1.j2 +++ b/roles/openshift_node/templates/node.yaml.v1.j2 @@ -1,5 +1,9 @@ allowDisabledDocker: false apiVersion: v1 +{% if openshift.common.version_gte_3_6 %} +dnsBindAddress: 127.0.0.1:53 +dnsRecursiveResolvConf: /etc/origin/node/resolv.conf +{% endif %} dnsDomain: {{ openshift.common.dns_domain }} {% if 'dns_ip' in openshift.node %} dnsIP: {{ openshift.node.dns_ip }} @@ -12,6 +16,21 @@ imageConfig: latest: false kind: NodeConfig kubeletArguments: {{ openshift.node.kubelet_args | default(None) | to_padded_yaml(level=1) }} +{% if openshift_use_crio | default(False) %} + container-runtime: + - remote + container-runtime-endpoint: + - /var/run/crio.sock + experimental-cri: + - 'true' + image-service-endpoint: + - /var/run/crio.sock + node-labels: + - router=true + - registry=true + runtime-request-timeout: + - 10m +{% endif %} {% if openshift.common.version_gte_3_3_or_1_3 | bool %} masterClientConnectionOverrides: acceptContentTypes: application/vnd.kubernetes.protobuf,application/json diff --git a/roles/openshift_node/templates/openshift.docker.node.dep.service b/roles/openshift_node/templates/openshift.docker.node.dep.service index 4c47f8c0d..8734e7443 100644 --- a/roles/openshift_node/templates/openshift.docker.node.dep.service +++ b/roles/openshift_node/templates/openshift.docker.node.dep.service @@ -3,7 +3,7 @@ Requires={{ openshift.docker.service_name }}.service After={{ openshift.docker.service_name }}.service PartOf={{ openshift.common.service_type }}-node.service Before={{ openshift.common.service_type }}-node.service - +{% if openshift_use_crio|default(false) %}Wants=cri-o.service{% endif %} [Service] ExecStart=/bin/bash -c "if [[ -f /usr/bin/docker-current ]]; then echo \"DOCKER_ADDTL_BIND_MOUNTS=--volume=/usr/bin/docker-current:/usr/bin/docker-current:ro --volume=/etc/sysconfig/docker:/etc/sysconfig/docker:ro\" > /etc/sysconfig/{{ openshift.common.service_type }}-node-dep; else echo \"#DOCKER_ADDTL_BIND_MOUNTS=\" > /etc/sysconfig/{{ openshift.common.service_type }}-node-dep; fi" diff --git a/roles/openshift_node/templates/openshift.docker.node.service b/roles/openshift_node/templates/openshift.docker.node.service index d89b64b06..639b6f6c8 100644 --- a/roles/openshift_node/templates/openshift.docker.node.service +++ b/roles/openshift_node/templates/openshift.docker.node.service @@ -12,14 +12,20 @@ After=ovs-vswitchd.service Wants={{ openshift.common.service_type }}-master.service Requires={{ openshift.common.service_type }}-node-dep.service After={{ openshift.common.service_type }}-node-dep.service +Requires=dnsmasq.service +After=dnsmasq.service [Service] EnvironmentFile=/etc/sysconfig/{{ openshift.common.service_type }}-node EnvironmentFile=/etc/sysconfig/{{ openshift.common.service_type }}-node-dep ExecStartPre=-/usr/bin/docker rm -f {{ openshift.common.service_type }}-node -ExecStart=/usr/bin/docker run --name {{ openshift.common.service_type }}-node --rm --privileged --net=host --pid=host --env-file=/etc/sysconfig/{{ openshift.common.service_type }}-node -v /:/rootfs:ro,rslave -e CONFIG_FILE=${CONFIG_FILE} -e OPTIONS=${OPTIONS} -e HOST=/rootfs -e HOST_ETC=/host-etc -v {{ openshift.common.data_dir }}:{{ openshift.common.data_dir }}{{ ':rslave' if openshift.docker.gte_1_10 | default(False) | bool else '' }} -v {{ openshift.common.config_base }}/node:{{ openshift.common.config_base }}/node {% if openshift_cloudprovider_kind | default('') != '' -%} -v {{ openshift.common.config_base }}/cloudprovider:{{ openshift.common.config_base}}/cloudprovider {% endif -%} -v /etc/localtime:/etc/localtime:ro -v /etc/machine-id:/etc/machine-id:ro -v /run:/run -v /sys:/sys:rw -v /sys/fs/cgroup:/sys/fs/cgroup:rw -v /usr/bin/docker:/usr/bin/docker:ro -v /var/lib/docker:/var/lib/docker -v /lib/modules:/lib/modules -v /etc/origin/openvswitch:/etc/openvswitch -v /etc/origin/sdn:/etc/openshift-sdn -v /var/lib/cni:/var/lib/cni -v /etc/systemd/system:/host-etc/systemd/system -v /var/log:/var/log -v /dev:/dev $DOCKER_ADDTL_BIND_MOUNTS {{ openshift.node.node_image }}:${IMAGE_VERSION} +ExecStartPre=/usr/bin/cp /etc/origin/node/node-dnsmasq.conf /etc/dnsmasq.d/ +ExecStartPre=/usr/bin/dbus-send --system --dest=uk.org.thekelleys.dnsmasq /uk/org/thekelleys/dnsmasq uk.org.thekelleys.SetDomainServers array:string:/in-addr.arpa/127.0.0.1,/{{ openshift.common.dns_domain }}/127.0.0.1 +ExecStart=/usr/bin/docker run --name {{ openshift.common.service_type }}-node --rm --privileged --net=host --pid=host --env-file=/etc/sysconfig/{{ openshift.common.service_type }}-node -v /:/rootfs:ro,rslave -e CONFIG_FILE=${CONFIG_FILE} -e OPTIONS=${OPTIONS} -e HOST=/rootfs -e HOST_ETC=/host-etc -v {{ openshift.common.data_dir }}:{{ openshift.common.data_dir }}{{ ':rslave' if openshift.docker.gte_1_10 | default(False) | bool else '' }} -v {{ openshift.common.config_base }}/node:{{ openshift.common.config_base }}/node {% if openshift_cloudprovider_kind | default('') != '' -%} -v {{ openshift.common.config_base }}/cloudprovider:{{ openshift.common.config_base}}/cloudprovider {% endif -%} -v /etc/localtime:/etc/localtime:ro -v /etc/machine-id:/etc/machine-id:ro -v /run:/run -v /sys:/sys:rw -v /sys/fs/cgroup:/sys/fs/cgroup:rw -v /usr/bin/docker:/usr/bin/docker:ro -v /var/lib/docker:/var/lib/docker -v /lib/modules:/lib/modules -v /etc/origin/openvswitch:/etc/openvswitch -v /etc/origin/sdn:/etc/openshift-sdn -v /var/lib/cni:/var/lib/cni -v /etc/systemd/system:/host-etc/systemd/system -v /var/log:/var/log -v /dev:/dev $DOCKER_ADDTL_BIND_MOUNTS -v /etc/pki:/etc/pki:ro {{ openshift.node.node_image }}:${IMAGE_VERSION} ExecStartPost=/usr/bin/sleep 10 ExecStop=/usr/bin/docker stop {{ openshift.common.service_type }}-node +ExecStopPost=/usr/bin/rm /etc/dnsmasq.d/node-dnsmasq.conf +ExecStopPost=/usr/bin/dbus-send --system --dest=uk.org.thekelleys.dnsmasq /uk/org/thekelleys/dnsmasq uk.org.thekelleys.SetDomainServers array:string: SyslogIdentifier={{ openshift.common.service_type }}-node Restart=always RestartSec=5s diff --git a/roles/openshift_node/templates/origin-node.service.j2 b/roles/openshift_node/templates/origin-node.service.j2 deleted file mode 100644 index 8047301e6..000000000 --- a/roles/openshift_node/templates/origin-node.service.j2 +++ /dev/null @@ -1,21 +0,0 @@ -[Unit] -Description=Origin Node -After={{ openshift.docker.service_name }}.service -Wants={{ openshift.docker.service_name }}.service -Documentation=https://github.com/openshift/origin - -[Service] -Type=notify -EnvironmentFile=/etc/sysconfig/origin-node -Environment=GOTRACEBACK=crash -ExecStart=/usr/bin/openshift start node --config=${CONFIG_FILE} $OPTIONS -LimitNOFILE=65536 -LimitCORE=infinity -WorkingDirectory=/var/lib/origin/ -SyslogIdentifier=origin-node -Restart=always -RestartSec=5s -OOMScoreAdjust=-999 - -[Install] -WantedBy=multi-user.target diff --git a/roles/openshift_node/templates/tuned/openshift-control-plane/tuned.conf b/roles/openshift_node/templates/tuned/openshift-control-plane/tuned.conf new file mode 100644 index 000000000..f22f21065 --- /dev/null +++ b/roles/openshift_node/templates/tuned/openshift-control-plane/tuned.conf @@ -0,0 +1,25 @@ +# +# tuned configuration +# + +[main] +summary=Optimize systems running OpenShift control plane +include=openshift + +[sysctl] +# ktune sysctl settings, maximizing i/o throughput +# +# Minimal preemption granularity for CPU-bound tasks: +# (default: 1 msec# (1 + ilog(ncpus)), units: nanoseconds) +kernel.sched_min_granularity_ns=10000000 + +# The total time the scheduler will consider a migrated process +# "cache hot" and thus less likely to be re-migrated +# (system default is 500000, i.e. 0.5 ms) +kernel.sched_migration_cost_ns=5000000 + +# SCHED_OTHER wake-up granularity. +# +# Preemption granularity when tasks wake up. Lower the value to improve +# wake-up latency and throughput for latency critical tasks. +kernel.sched_wakeup_granularity_ns = 4000000 diff --git a/roles/openshift_node/templates/tuned/openshift-node/tuned.conf b/roles/openshift_node/templates/tuned/openshift-node/tuned.conf new file mode 100644 index 000000000..78c7d19c9 --- /dev/null +++ b/roles/openshift_node/templates/tuned/openshift-node/tuned.conf @@ -0,0 +1,10 @@ +# +# tuned configuration +# + +[main] +summary=Optimize systems running OpenShift nodes +include=openshift + +[sysctl] +net.ipv4.tcp_fastopen=3 diff --git a/roles/openshift_node/templates/tuned/openshift/tuned.conf b/roles/openshift_node/templates/tuned/openshift/tuned.conf new file mode 100644 index 000000000..68ac5dadb --- /dev/null +++ b/roles/openshift_node/templates/tuned/openshift/tuned.conf @@ -0,0 +1,24 @@ +# +# tuned configuration +# + +[main] +summary=Optimize systems running OpenShift (parent profile) +include=${f:virt_check:{{ openshift_tuned_guest_profile }}:throughput-performance} + +[selinux] +avc_cache_threshold=65536 + +[net] +nf_conntrack_hashsize=131072 + +[sysctl] +kernel.pid_max=131072 +net.netfilter.nf_conntrack_max=1048576 +fs.inotify.max_user_watches=65536 +net.ipv4.neigh.default.gc_thresh1=8192 +net.ipv4.neigh.default.gc_thresh2=32768 +net.ipv4.neigh.default.gc_thresh3=65536 +net.ipv6.neigh.default.gc_thresh1=8192 +net.ipv6.neigh.default.gc_thresh2=32768 +net.ipv6.neigh.default.gc_thresh3=65536 diff --git a/roles/openshift_node/templates/tuned/recommend.conf b/roles/openshift_node/templates/tuned/recommend.conf new file mode 100644 index 000000000..5fa765798 --- /dev/null +++ b/roles/openshift_node/templates/tuned/recommend.conf @@ -0,0 +1,8 @@ +[openshift-node] +/etc/origin/node/node-config.yaml=.*region=primary + +[openshift-control-plane,master] +/etc/origin/master/master-config.yaml=.* + +[openshift-control-plane,node] +/etc/origin/node/node-config.yaml=.*region=infra |