Elasticsearch is packaged into the Mission Control installation. The following describes the different variations for each distribution package type:
|Docker||Elasticsearch for Docker is added as a Docker container to the Mission Control docker-compose project. This installs version 6.1.1.|
|Elasticsearch 6.1.1 is included in the Mission Control Debian project. Linux service files are added.|
|RPM||Elasticsearch 6.1.1 is included in the Mission Control RPM project. Linux service files are added.|
After Mission Control is fully installed, the Elasticsearch data files can be found in the following locations:
Elasticsearch uses the following communication ports:
There are two aliases to store and retrieve data:
|active_insight_data||Used point to active index to push data.|
|search_insight_data||Used to search and retrieve data|
On installation, these aliases and indices of format active_insight_data_timestamp* are created.
Mission Control is pre-configured to periodically cleanup indexes to keep data for a period of one year.
As the amount of historical data collected by Mission Control accumulates, you may want to scale the Elasticsearch database used to store that data to maintain performance and responsiveness.
To scale a running Elasticsearch instance to a two-node cluster, follow the steps below:
Add the following settings to your startup scripts. In the case of a Docker installation, add them to the
docker-compose file, for a non-Docker installation, add them to the
discovery.zen.minimum_master_nodes=2 #(The recommendation is for (N/2+1), where N is the number of eligible master nodes). node.master=true node.data=true discovery.zen.ping.unicast.hosts=<Published IP address of each node> network.publish_host=<IP address to be published for the node> node.name=<Node name>
First time you run Elasticsearch as a two-node cluster, you need to make sure to restart your first node (the one that already contains data) before adding the second one. Starting up the new node first could cause index mapping templates to be deleted.
Once your cluster is set up, Mission Control will automatically detect the new node at runtime and start sending it requests.
If your Elasticsearch cluster is behind a load balancer, you need to add the following environment variables to the
insight_server service to provide it with the load balancer read and write URLs: