Bootstrapping a clusteredit

Starting an Elasticsearch cluster for the very first time requires the initial set of master-eligible nodes to be explicitly defined on one or more of the master-eligible nodes in the cluster. This is known as cluster bootstrapping. This is only required the first time a cluster starts up: nodes that have already joined a cluster store this information in their data folder for use in a full cluster restart, and freshly-started nodes that are joining a running cluster obtain this information from the cluster’s elected master.

The initial set of master-eligible nodes is defined in the cluster.initial_master_nodes setting. This should be set to a list containing one of the following items for each master-eligible node:

  • The node name of the node.
  • The node’s hostname if node.name is not set, because node.name defaults to the node’s hostname. You must use either the fully-qualified hostname or the bare hostname depending on your system configuration.
  • The IP address of the node’s publish address, if it is not possible to use the node.name of the node. This is normally the IP address to which network.host resolves but this can be overridden.
  • The IP address and port of the node’s publish address, in the form IP:PORT, if it is not possible to use the node.name of the node and there are multiple nodes sharing a single IP address.

When you start a master-eligible node, you can provide this setting on the command line or in the elasticsearch.yml file. After the cluster has formed, this setting is no longer required. It should not be set for master-ineligible nodes, master-eligible nodes joining an existing cluster, or cluster restarts.

It is technically sufficient to set cluster.initial_master_nodes on a single master-eligible node in the cluster, and only to mention that single node in the setting’s value, but this provides no fault tolerance before the cluster has fully formed. It is therefore better to bootstrap using at least three master-eligible nodes, each with a cluster.initial_master_nodes setting containing all three nodes.

You must set cluster.initial_master_nodes to the same list of nodes on each node on which it is set in order to be sure that only a single cluster forms during bootstrapping and therefore to avoid the risk of data loss.

For a cluster with 3 master-eligible nodes (with node names master-a, master-b and master-c) the configuration will look as follows:

cluster.initial_master_nodes:
  - master-a
  - master-b
  - master-c

Like all node settings, it is also possible to specify the initial set of master nodes on the command-line that is used to start Elasticsearch:

$ bin/elasticsearch -Ecluster.initial_master_nodes=master-a,master-b,master-c

The node names used in the cluster.initial_master_nodes list must exactly match the node.name properties of the nodes. By default the node name is set to the machine’s hostname which may or may not be fully-qualified depending on your system configuration. If each node name is a fully-qualified domain name such as master-a.example.com then you must use fully-qualified domain names in the cluster.initial_master_nodes list too; conversely if your node names are bare hostnames (without the .example.com suffix) then you must use bare hostnames in the cluster.initial_master_nodes list. If you use a mix of fully-qualifed and bare hostnames, or there is some other mismatch between node.name and cluster.initial_master_nodes, then the cluster will not form successfully and you will see log messages like the following.

[master-a.example.com] master not discovered yet, this node has
not previously joined a bootstrapped (v7+) cluster, and this
node must discover master-eligible nodes [master-a, master-b] to
bootstrap a cluster: have discovered [{master-b.example.com}{...

This message shows the node names master-a.example.com and master-b.example.com as well as the cluster.initial_master_nodes entries master-a and master-b, and it is clear from this message that they do not match exactly.

Choosing a cluster nameedit

The cluster.name setting enables you to create multiple clusters which are separated from each other. Nodes verify that they agree on their cluster name when they first connect to each other, and Elasticsearch will only form a cluster from nodes that all have the same cluster name. The default value for the cluster name is elasticsearch, but it is recommended to change this to reflect the logical name of the cluster.

Auto-bootstrapping in development modeedit

If the cluster is running with a completely default configuration then it will automatically bootstrap a cluster based on the nodes that could be discovered to be running on the same host within a short time after startup. This means that by default it is possible to start up several nodes on a single machine and have them automatically form a cluster which is very useful for development environments and experimentation. However, since nodes may not always successfully discover each other quickly enough this automatic bootstrapping cannot be relied upon and cannot be used in production deployments.

If any of the following settings are configured then auto-bootstrapping will not take place, and you must configure cluster.initial_master_nodes as described in the section on cluster bootstrapping:

  • discovery.seed_providers
  • discovery.seed_hosts
  • cluster.initial_master_nodes

If you start an Elasticsearch node without configuring these settings then it will start up in development mode and auto-bootstrap itself into a new cluster. If you start some Elasticsearch nodes on different hosts then by default they will not discover each other and will form a different cluster on each host. Elasticsearch will not merge separate clusters together after they have formed, even if you subsequently try and configure all the nodes into a single cluster. This is because there is no way to merge these separate clusters together without a risk of data loss. You can tell that you have formed separate clusters by checking the cluster UUID reported by GET / on each node. If you intended to form a single cluster then you should start again:

  • Take a snapshot of each of the single-host clusters if you do not want to lose any data that they hold. Note that each cluster must use its own snapshot repository.
  • Shut down all the nodes.
  • Completely wipe each node by deleting the contents of their data folders.
  • Configure cluster.initial_master_nodes as described above.
  • Restart all the nodes and verify that they have formed a single cluster.
  • Restore any snapshots as required.