Refer to the {cloud}/ec-adding-plugins.html[{ess} documentation] for information about managing plugins on {ecloud}.
Use the elasticsearch-plugin
command line tool to install, list, and remove plugins. It is
located in the $ES_HOME/bin
directory by default but it may be in a
different location depending on which Elasticsearch package you installed:
-
{ref}/targz.html#targz-layout[Directory layout of
.tar.gz
archives] -
{ref}/zip-windows.html#windows-layout[Directory layout of Windows
.zip
archives] -
{ref}/deb.html#deb-layout[Directory layout of Debian package]
-
{ref}/rpm.html#rpm-layout[Directory layout of RPM]
Run the following command to get usage instructions:
sudo bin/elasticsearch-plugin -h
Important
|
Running as root
If Elasticsearch was installed using the deb or rpm package then run
|
If you run {es} using Docker, you can manage plugins using a configuration file.
The documentation for each plugin usually includes specific installation instructions for that plugin, but below we document the various available options:
Core Elasticsearch plugins can be installed as follows:
sudo bin/elasticsearch-plugin install [plugin_name]
For instance, to install the core ICU plugin, just run the following command:
sudo bin/elasticsearch-plugin install analysis-icu
This command will install the version of the plugin that matches your Elasticsearch version and also show a progress bar while downloading.
A plugin can also be downloaded directly from a custom location by specifying the URL:
sudo bin/elasticsearch-plugin install [url] (1)
-
must be a valid URL, the plugin name is determined from its descriptor.
- Unix
-
To install a plugin from your local file system at
/path/to/plugin.zip
, you could run:sudo bin/elasticsearch-plugin install file:///path/to/plugin.zip
- Windows
-
To install a plugin from your local file system at
C:\path\to\plugin.zip
, you could run:bin\elasticsearch-plugin install file:///C:/path/to/plugin.zip
NoteAny path that contains spaces must be wrapped in quotes! NoteIf you are installing a plugin from the filesystem the plugin distribution must not be contained in the plugins
directory for the node that you are installing the plugin to or installation will fail. - HTTP
-
To install a plugin from an HTTP URL:
sudo bin/elasticsearch-plugin install https://some.domain/path/to/plugin.zip
The plugin script will refuse to talk to an HTTPS URL with an untrusted certificate. To use a self-signed HTTPS cert, you will need to add the CA cert to a local Java truststore and pass the location to the script as follows:
sudo CLI_JAVA_OPTS="-Djavax.net.ssl.trustStore=/path/to/trustStore.jks" bin/elasticsearch-plugin install https://host/plugin.zip
Multiple plugins can be installed in one invocation as follows:
sudo bin/elasticsearch-plugin install [plugin_id] [plugin_id] ... [plugin_id]
Each plugin_id
can be any valid form for installing a single plugin (e.g., the
name of a core plugin, or a custom URL).
For instance, to install the core ICU plugin, run the following command:
sudo bin/elasticsearch-plugin install analysis-icu
This command will install the versions of the plugins that matches your Elasticsearch version. The installation will be treated as a transaction, so that all the plugins will be installed, or none of the plugins will be installed if any installation fails.
If you rely on some plugins, you can define mandatory plugins by adding
plugin.mandatory
setting to the config/elasticsearch.yml
file, for
example:
plugin.mandatory: analysis-icu,lang-js
For safety reasons, a node will not start if it is missing a mandatory plugin.
A list of the currently loaded plugins can be retrieved with the list
option:
sudo bin/elasticsearch-plugin list
Alternatively, use the {ref}/cluster-nodes-info.html[node-info API] to find out which plugins are installed on each node in the cluster
Plugins can be removed manually, by deleting the appropriate directory under
plugins/
, or using the public script:
sudo bin/elasticsearch-plugin remove [pluginname]
After a Java plugin has been removed, you will need to restart the node to complete the removal process.
By default, plugin configuration files (if any) are preserved on disk; this is
so that configuration is not lost while upgrading a plugin. If you wish to
purge the configuration files while removing a plugin, use -p
or --purge
.
This can option can be used after a plugin is removed to remove any lingering
configuration files.
Multiple plugins can be removed in one invocation as follows:
sudo bin/elasticsearch-plugin remove [pluginname] [pluginname] ... [pluginname]
Except for text analysis plugins that are created using the stable plugin API, plugins are built for a specific version of {es}, and must be reinstalled each time {es} is updated.
sudo bin/elasticsearch-plugin remove [pluginname]
sudo bin/elasticsearch-plugin install [pluginname]
The plugin
scripts supports a number of other command line parameters:
The --verbose
parameter outputs more debug information, while the --silent
parameter turns off all output including the progress bar. The script may
return the following exit codes:
0
|
everything was OK |
64
|
unknown command or incorrect option parameter |
74
|
IO error |
70
|
any other error |
Certain plugins require more privileges than those provided by default in core Elasticsearch. These plugins will list the required privileges and ask the user for confirmation before continuing with installation.
When running the plugin install script from another program (e.g. install
automation scripts), the plugin script should detect that it is not being
called from the console and skip the confirmation response, automatically
granting all requested permissions. If console detection fails, then batch
mode can be forced by specifying -b
or --batch
as follows:
sudo bin/elasticsearch-plugin install --batch [pluginname]
If your elasticsearch.yml
config file is in a custom location, you will need
to specify the path to the config file when using the plugin
script. You
can do this as follows:
sudo ES_PATH_CONF=/path/to/conf/dir bin/elasticsearch-plugin install <plugin name>
To install a plugin via a proxy, you can add the proxy details to the
CLI_JAVA_OPTS
environment variable with the Java settings http.proxyHost
and http.proxyPort
(or https.proxyHost
and https.proxyPort
):
sudo CLI_JAVA_OPTS="-Dhttp.proxyHost=host_name -Dhttp.proxyPort=port_number -Dhttps.proxyHost=host_name -Dhttps.proxyPort=https_port_number" bin/elasticsearch-plugin install analysis-icu
Or on Windows:
set CLI_JAVA_OPTS="-Dhttp.proxyHost=host_name -Dhttp.proxyPort=port_number -Dhttps.proxyHost=host_name -Dhttps.proxyPort=https_port_number"
bin\elasticsearch-plugin install analysis-icu
The default location of the plugins
directory depends on which package you install:
-
{ref}/targz.html#targz-layout[Directory layout of
.tar.gz
archives] -
{ref}/zip-windows.html#windows-layout[Directory layout of Windows
.zip
archives] -
{ref}/deb.html#deb-layout[Directory layout of Debian package]
-
{ref}/rpm.html#rpm-layout[Directory layout of RPM]
Important
|
Docker only
This feature is only available for official {es} Docker images. Other {es} distributions will not start with a plugin configuration file. |
If you run {es} using Docker, you can manage plugins using a declarative configuration file. When {es} starts up, it will compare the plugins in the file with those that are currently installed, and add or remove plugins as required. {es} will also upgrade official plugins when you upgrade {es} itself.
The file is called elasticsearch-plugins.yml
, and must be placed in the
Elasticsearch configuration directory, alongside elasticsearch.yml
. Here
is an example:
plugins:
- id: analysis-icu
- id: repository-azure
- id: custom-mapper
location: https://example.com/archive/custom-mapper-1.0.0.zip
This example installs the official analysis-icu
and
repository-azure
plugins, and one unofficial plugin. Every plugin must provide
an id
. Unofficial plugins must also provide a location
. This is
typically a URL, but Maven coordinates are also supported. The downloaded
plugin’s name must match the ID in the configuration file.
While {es} will respect the standard Java proxy system properties when downloading plugins, you can also configure an HTTP proxy to use explicitly in the configuration file. For example:
plugins:
- id: custom-mapper
location: https://example.com/archive/custom-mapper-1.0.0.zip
proxy: proxy.example.com:8443