elastic-role-playbook's Introduction

to deploy cluster

ansible-playbook -i inventories/vagrant --ask-become-pass cluster-prep.yaml #deploys packages, configure hostnames and disables firewall
ansible-playbook -i inventories/vagrant --ask-become-pass deploy_cluster_uri.yaml  #deploys cluster, users and roles for fms, creates default index temlate, adds s3 repo
ansible-playbook -i inventories/vagrant --ask-become-pass deploy_kibana.yaml #deploys kibana, creates fms space.

testing repo with minio (tested version elastic 7.7.0):

Deploy minio in docker in masternode, and make sure port is accessible. Also install mc client as explained in minio docs.

mc config host add myminio http://127.0.0.1:9000 minioadmin minioadmin

```
mc mb myminio/elasticsearch
```

export S3_SECRET_KEY=minioadmin; export S3_ACCESS_KEY=minioadmin

echo $S3_SECRET_KEY | /usr/share/elasticsearch/bin/elasticsearch-keystore add --stdin s3.client.default.secret_key

echo $S3_ACCESS_KEY | /usr/share/elasticsearch/bin/elasticsearch-keystore add --stdin s3.client.default.access_key

elasticsearch.yml: 
s3.client.default.endpoint: "http://127.0.0.1:9000"
s3.client.default.protocol: "http"

Create a json file with the following contents:

{ "type": "s3", "settings": { "bucket": "elasticsearch" } }
elasticsearch defaults to amazon s3, so thats why you need to add those settings in elasticsearch.yml
in 7.7 version (probably also in 7.x) keys to access s3 repo MUST be stored in the keystore, so there you have the commands to add them.
in the json to create the repo, you don't need to add anything else than the bucket name, and the default path.
DO A RESTART OF ELASTIC IN ALL NODES!!!
Do not freak out when you see amazon s3 exceptions when troubleshooting this, just make sure to have this config.
The key for this to work with minio is make sure setting up the exact same endpoint in minio that you setup in the mc command.