Overview
Cluster name
It is important to change the name of the cluster in elasticsearch.yml to avoid Elasticsearch nodes joining the wrong cluster. This is particularly important when development, staging and production environments can find themselves on the same network.
How to prevent it from happening
If you want to change the name of the cluster, then you need to modify the setting in elasticsearch.yml and perform a rolling restart:
cluster.name: myapp-prod
Remember that changing the cluster name will also change the default name of your Elasticsearch log files.
Data path
If you have installed Elasticsearch using zip or tar.gz files, then by default the data directory is a subdirectory of the Elasticsearch home directory. This creates a high risk of potentially deleting your data when upgrading Elasticsearch.
How to prevent it from happening
Make sure you set the data and logs directories in elasticsearch.yml to a path separated from your Elasticsearch program files:
path: logs: /var/log/elasticsearch data: /var/data/elasticsearch
These directories must be writable by the Elasticsearch user.
If you have already created data in the default path, and you want to move the directory, then you should follow the steps in the procedure below.
How to change the data path:
- Double check to ensure you have a recent snapshot of all indices on the node
- Temporarily stop shard relocation using:
curl -XPUT localhost:9200/_cluster/settings -d ‘{
“transient” : {
“cluster.routing.allocation.enable” : “none”
}
}’ - Stop the Elasticsearch node
- Move the entire data directory to its new location
- Modify the path in elasticsearch.yml
- Start the Elasticsearch node
Please be extra careful when taking the above steps and make sure they fit your system, as misusing them can lead to loss of production data.
It is recommended to use RPM or Debian packages to avoid this, and other installation issues. RPM and Debian packages by default store data separately from program files.
