Page tree
Skip to end of metadata
Go to start of metadata

Use this command to change cluster-level settings for Indexers present in your environment.

These settings can help you customize the cluster behavior to better suit your particular deployment scenario

Note

The default values for cluster-level settings are optimum and changing them can impact performance. BMC recommends that you do not change the default values unless you are instructed to by BMC Customer Support.


 

 

To understand the syntax and examples, see:

Syntax

Note

Before you run the command, note the following points:

configureclustersettings -d <hostname> -p <portNumber> -u <username> -w <password> [-l <locale>] [-s] -ops <operatioType>

  • Angle brackets (< >) indicate a variable value that needs to be user defined.

  • Square brackets ([ ]) indicate optional parameters.

The following table describes the parameters used in the syntax:

ParameterDescription
-ops
        

Type of operation that you want to perform.

Depending on the type of operation that you want to perform, the value can be list or modify. For more information about the settings, see Cluster-level settings.

The supported operations are explained as follows:

List operations
To...Specify the value
List details of particular settings (property names specified).

-ops list -k <keysToRead>

In the preceding syntax, <keysToRead> refers to comma-separated list of property names for which you want to list details.

List details of all the settings recently modified.
-ops list 
Modification operations
To...Specify the value

Modify details of the setting.

The value can be specified in one of the following ways:

  • Specify a list of keys (or property names) with new values.
  • Provide details of a file containing the key=value pairs with new values.

 modify -k <keysToUpdate>

In the preceding syntax, <keysToUpdate> refers to a comma-separated list of property names for which you want to update details.

The keys must be specified in the format, <propertyName1>=<propertyNewValue1>,<propertyName1>=<propertyNewValue1>.

Example: -ops modify -k cluster.routing.allocation.disk.watermark.low=10gb,
cluster.routing.allocation.disk.watermark.high=5gb

modify -f <fileDetails>

In the preceding syntax, <fileDetails> refers to the complete path to the file containing details of the properties with new values.
The properties in the file must be added in the format, key=value. Each property must begin on a new line. The file must be saved with a .properties file extension.

-d
Host name of the Console Server 
-p
Port number of the Console Server. 
-u
User name to use to connect with the Console Server. 
-v
(Optional) Version of the IT Data Analytics product. 
-w
Password corresponding to the user name to use to connect with the Console Server. 
-l

(Optional) Indicates the locale in which you want to display messages when the command is run.

You can set one of the following locales:

  • English
  • Simplified Chinese
  • German
  • French

By default, the messages are displayed in the locale of the system on which the CLI is run.

The value can be set to one of the following depending on the locale you choose:

  • English: en
  • Simplified Chinese: chs
  • German: de
  • French: fr

Example: If you want to set the locale as Simplified Chinese, you need to specify, l=chs.

 
-s
(Optional) Indicates an HTTPS connection. 

Cluster-level settings

The following table provides definitions of the cluster-level settings that you can change by running this command.

Cluster-level settingDescription
discovery.zen.publish_timeout

Determines the timeout period used by the master node to wait for the responses from the peer nodes before moving on to the next update job.

 The master node performs cluster state updates and publishes the updated state to all the other nodes in the cluster. Each node updates its own state and responds to the master.

Default value: 30 seconds

Involves cluster restart: No

cluster.routing.allocation.disk.threshold_enabled

Decides whether during shard allocation to a node, the amount of disk space available needs to be considered or not. When enabled, the shard allocation takes into account two watermark properties: high and low.

This setting can help rebalance shards by reallocating them from nodes with full disks to nodes with relatively lesser full disks.

Default value: true

Involves cluster restart: No

cluster.routing.allocation.disk.watermark.low

Controls the low watermark for disk usage.

A low watermark indicates that new shards will not be allocated to a node after the disk usage crosses 85% (default value). This means after the 85% point is crossed, the new shards will be allocated on other nodes with free disk space.

This value can be set as a percentage value or an absolute byte value (for example 500mb). Setting the value in bytes might be a more effective way of stopping shards from getting allocated to a node with less than the configured amount of space available.

Note that percentage values refer to used disk space and byte values refer to free disk space.

Default value: 85%

Involves cluster restart: No

cluster.routing.allocation.disk.watermark.high

Controls the high watermark for disk usage.

A high watermark indicates that shards existing on a node will be relocated to other nodes if the node disk usage rises above 90% (default value).

This value can be set as a percentage value or an absolute byte value (for example 500mb). Setting the value in bytes might be a more effective way of relocating shards from a node, when less than the configured amount of disk space is available on that node. 

Note that percentage values refer to used disk space and byte values refer to free disk space (similar to the low watermark setting).

Default value: 90%

Involves cluster restart: No

gateway.expected_nodes

Determines the expected number of master or peer Indexers that must be present in the cluster for the recover to start immediately.

Recovery of the local shards begins as soon as the expected number of nodes join the cluster.

Setting this property defaults the gateway.recover_after_time property to five minutes. Until the expected number of nodes are not started, gateway.recover_after_time determines the amount of time to wait before recovering the cluster. However, as soon as the expected number of nodes are started, the gateway.recover_after_time property is ignored.

Default value: 0

Involves cluster restart: No

gateway.recover_after_nodes

Determines the minimum number of nodes (master or peers) that must be already started before the cluster recovery begins.

Each time the cluster is restarted, a recovery process (gateway) is used to load all the relevant information about the cluster state (cluster metadata including indexes with their index-level settings and type mappings). With this setting, you can configure the recovery process to control when the cluster data must be recovered.

Ideally, this value must be equal to the number of nodes in the cluster. You might want the cluster state to be recovered only after all the nodes are started. However, at a minimum, this value must be more than half of the number of nodes available in the cluster.

Involves cluster restart: Yes

gateway.recover_after_time

Determines the waiting time before the recovery process must start and after the conditions for recovery are met.

Default value: 5 m (if the gateway.expected_nodes property is set)

Involves cluster restart: Yes

Example inputs

The following example provides the input for listing details of all the Indexer cluster settings.

configureclustersettings -u admin -w admin12345 -d win-pun-01 -p 9797 -ops list