21f1f28ab3
Fixes for this patchset: - split out elk-openstack-client.yml to match what's done elsewhere Fixes for patchset #11: - split out filebeat into separate role for openstack clients - update README.md to use elk-openstack-client.yml for this purpose - cleanup filebeat.yml.j2 to use correct syntax (no need for " anymore) Fixes for patchset #10: - add SELinux boolean "httpd_can_network_connect" - add libsemanage-python package dependency for booleans Fixes for patchset #9: - fix for RHEL7 clients, we need to specify remote EPEL rpm - RHEL7 clients need rpm_key module to import EPEL GPG key - switch to using uri module instead of curl for checking elasticsearch indices - add python-httplib2 dependency (needed for uri module) - use curl -XPOST instead of PUT for filebeat index template in elasticsearch Fixes from patchset #7 - remove unneeded rpm usage, switch to yum module - add logic to heapsize tuning so systems > 64G of memory will never exceed the 32G recommended heapsize - logic fix for prepopulating local logs into logstash - remove elasticsearch.yml, rpm provides this and we're not customizing it yet Fixes from patchset #6: - use yum repo Ansible module where we can - remove unecessary EPEL installation (only nginx needs it) - disable EPEL repo after installation to avoid OpenStack breakage This adds: (ELK Server) - Automated ELK stack deployment - SSL client generation - Heap size tuning (1/2 of available memory) - Firewall port additions (depending on active or not) - Supports either firewalld or iptables-services - Additional upstream Filebeat Kibana dashboards (ELK Client) - Sets up filebeat with appropriate SSL certificates - utilizes both hostnames and SubjectAltName support (for environments without DNS services). (Usage) ansible-playbook -i hosts install/elk.yml ansible-playbook -i hosts install/elk-client.yml --extra-vars 'elk_server=X.X.X.X' Change-Id: Iee29f985e0bbcdf706ad869f132d4c0f1593a6b6
51 lines
1.6 KiB
YAML
51 lines
1.6 KiB
YAML
---
|
|
#
|
|
# Install/run elasticsearch for browbeat
|
|
#
|
|
|
|
- name: Copy elasticsearch yum repo file
|
|
copy:
|
|
src=elasticsearch.repo
|
|
dest=/etc/yum.repos.d/elasticsearch.repo
|
|
owner=root
|
|
group=root
|
|
mode=0644
|
|
become: true
|
|
|
|
- name: Install elasticsearch and openjdk
|
|
yum: name={{ item }} state=present
|
|
become: true
|
|
with_items:
|
|
- elasticsearch
|
|
- java-openjdk-headless
|
|
|
|
- name: Check if system memory is greater than 64G
|
|
debug: msg="System memory is {{ansible_memory_mb.real.total | int}} so setting heapsize to 32G upper limit"
|
|
when: ansible_memory_mb.real.total|int >= 65536
|
|
|
|
- name: Apply heapsize tuning for systems with greater than 64G memory
|
|
lineinfile: dest=/usr/share/elasticsearch/bin/elasticsearch.in.sh \
|
|
line="ES_HEAP_SIZE=32g" insertafter="^ES_CLASSPATH="
|
|
when: ansible_memory_mb.real.total|int >= 65536
|
|
register: elasticsearch_updated
|
|
|
|
- name: Print extended documentation for heapsize tuning
|
|
debug: msg="Refer to https://www.elastic.co/guide/en/elasticsearch/guide/current/_limiting_memory_usage.html"
|
|
when: ansible_memory_mb.real.total|int >= 65536
|
|
|
|
- name: Update elasticsearch startup with heap size
|
|
become: true
|
|
lineinfile: dest=/usr/share/elasticsearch/bin/elasticsearch.in.sh \
|
|
line="ES_HEAP_SIZE={{ (ansible_memory_mb.real.total / 2) | int }}m" insertafter="^ES_CLASSPATH="
|
|
when: ansible_memory_mb.real.total|int < 65536
|
|
register: elasticsearch_updated
|
|
|
|
- name: Start elasticsearch service
|
|
command: systemctl start elasticsearch.service
|
|
ignore_errors: true
|
|
when: elasticsearch_updated != 0
|
|
|
|
- name: Setup elasticsearch service
|
|
service: name=elasticsearch state=started enabled=true
|
|
become: true
|