Use the following to get a running instance of the Elastic Stack:
r10k puppetfile install
vagrant up --provision
This is an early version of the test setup and needs more work.
Most of the vagrant and puppet magic was inspired by the Vagrant boxes for the Icinga 2 book.
This is by no means a best-practice production setup.
This environment was built for:
- Having a test environment while updateing the slides for the Netways Elastic Stack Training. This is not used during the trainings because installation and configuration is part of them.
- Having a dev / test environment for the check_logstash monitoring plugin.
- Having an easy to use environment during support sessions
- Giving other users a quick and easy way to try Elastic Stack 6
Maybe this project will mature into a best-practice showcase if people are interested.
The Puppet code is not as nice as it could be. e.g. there is heavy usage of file_line instead of templates. This is due to the fact that it should show what has to be changed from a default installation to get a working setup and maybe later get a best-practice setup.
The Redis is mostly used to show the power of the multipipeline feature. You don't need it for a setupt with only beats but you will want it if you collect SNMP Traps, Syslog via UDP or anything else that can't deal with backpressure.
You can run a single instance by issueing the following command:
vagrant up elastic01 --provision
This capability is the main reason why some cluster tweaks in elasticsearch.yml are missing.
You can change the configuration in the logstash-indexer directory. This directory is mounted into the boxes and Logstash will automatically reload new configuration after changes.
You can use vagrant ssh elastic01 to log in to the main node.
The following ports are forwarded and you can use them for connecting via localhost.
- 5601 : kibana on elastic01
- 9200 : REST API of Elasticsearch on elastic01
- 9600 : REAST node API of Logstash on elastic01