Elasticsearch How to Upgrade from Version 5 to Version 6 in Elasticsearch

By Opster Expert Team

Updated: Jan 28, 2024

| 6 min read

Quick links

• Upgrade Elasticsearch 5.6.16 to Elasticsearch 6.8.23
• Preparing to upgrade Elasticsearch nodes
  1. Check the deprecation log
  2. Review the breaking changes
  3. Check the Elasticsearch plugins compatibility
     • X Pack Plugins
  4. Set up a test environment
  5. Take a backup and snapshots from your data
     • Reindex any Version 2 indices
• How to upgrade nodes in a rolling upgrade

Upgrade Elasticsearch 5.6.16 to Elasticsearch 6.8.23

In this article, we will focus on upgrading Elasticsearch 5.6.16 to Elasticsearch 6.8.23. For more information on version upgrades in general, the different methods available, and everything you need to know prior to initiating the upgrade, you can see our complete guide here: How to Upgrade Elasticsearch Versions.  

Rolling upgrades are only supported from Elasticsearch version 5.6.16 to Elasticsearch version 6.8.23. To upgrade from earlier versions of 5.6 to version 6.8.23, you need a full cluster restart or you must upgrade first to version 5.6.16 before upgrading to version 6.

Preparing to upgrade Elasticsearch nodes

Elasticsearch nodes cannot be downgraded after upgrading. Before starting the upgrade process you should review the following: 

1. Check the deprecation log 

You should read and resolve any issues highlighted in the deprecation log. These logs are usually located in: 

/var/log/elasticsearch/Your-Cluster-Name_deprecation.log

2. Review the breaking changes

Breaking changes documentation is published with each new version of Elasticsearch to make you aware of any functionality that may change or disappear. You should always check that documentation to ensure that none of those settings, configurations, or mappings are being used in your setup. The main things that could be affected are:

• Node configuration (elasticsearch.yml)
• Index mappings and templates
• Cluster settings
• Code used in Elasticsearch clients (applications and indexing scripts)

The most important breaking changes for version 6 are:

• Users are no longer allowed multiple index types per index – just one type is permitted per index
• The Groovy and Python scripting languages have been removed and replaced with Painless scripting

3. Check Elasticsearch plugins’ compatibility

If you are using any Elasticsearch plugins, you should check the availability and compatibility of those plugins with the new version.

X-Pack plugins

The changes made to the X-Pack plugins between version 5.6 and version 6.8 mean that a number of configurations and the internal indices used for version 5 are not compatible with version 6.

So, if you are running X-Pack, you should:

• Use the Upgrade Assistant in Kibana to identify and resolve upgrade issues.
• Upgrade the internal X-Pack indices using X-Pack 5.6 Reindex Helper in Kibana.
• Enable TLS because it is required in order to run with security enabled in version 6.

The following APIs can also be used as alternatives to the Kibana Upgrade Assistant:

To see any settings that have been deprecated and that should be fixed before upgrading:

/_xpack/migration/deprecations

To see a list of indices that need to be reindexed before upgrading to 6.8.23:

/_xpack/migration/assistance

To upgrade the X-Pack and security indices to a format that is compatible with version 6:

/_xpack/migration/upgrade

If you are NOT running X-Pack, then you will need to disable the security features in elasticsearch.yml and kibana.yml when upgrading to version 6:

xpack.security.enabled: false

4. Set up a test environment 

You should first try the upgrade process in a test or staging environment in order to check and resolve all issues before upgrading your production cluster. This is especially important if you are running X-Pack.

5. Take a backup and snapshots from your data  

Remember that it is not possible to downgrade an Elasticsearch node, so the only way to practically reverse a failed upgrade is to create a new cluster with the old version and recover your data from snapshots. Therefore, it is essential to take snapshots of all Elasticsearch indices before starting the upgrade process.

As an extra precaution, it is also advisable to export any Kibana visualizations to file.

Reindex any version 2 indices

Elasticsearch 6 can only support the indices created in version 5 or 6 , so if you have any indices that were created in version 1 or 2.x.x, you must reindex them. The upgraded node will not join the cluster if you have any version 1 or 2 indices.  

If you have X-Pack installed, the Upgrade Assistant will show you the indices you need to reindex. However, if you don’t have X-Pack, then you can check the Elasticsearch index version like this:

GET my-new-index-1/_settings

Result:

{
  "my-new-index-1" : {
    "settings" : {
      "index" : {
    			      .
.
.
        "version" : {
          "created" : "6082399"
        },
      }
    }
  }
}

To interpret the “created” field:

“Created”:  “XXYYZZAA”
// - XX is major version ⇒ here is major version 6, do not consider the zero
// - YY is minor version ⇒ here is minor version 08
// - ZZ is revision     ⇒ here is revision 23

So, based on the created field, you can see the index version is 6.8.23.

Upgrading nodes in a rolling upgrade

The process for upgrading your nodes is as follows, with all NON-master-eligible nodes upgraded first.

To stop a node:

systemctl stop elasticsearch

To disable shard allocation, run:

PUT _cluster/settings
{
  "persistent": {
    "cluster.routing.allocation.enable": "primaries"
  }
}

To remove plugins, e.g., X-Pack:

bin/elasticsearch-plugin remove x-pack

To reload the daemon after upgrading:

systemctl   daemon-reload
systemctl  enable elasticsearch

To re-enable shard allocation, run:

PUT _cluster/settings
{
  "persistent": {
    "cluster.routing.allocation.enable": null
  }
}

To get cluster status and see how many nodes are in the cluster, use:

GET _cluster/health