Updating number_of_replicas to for indices – How to solve related issues

Average Read Time

2 Mins

Updating number_of_replicas to for indices – How to solve related issues

Opster Team

Jan-20, Version: 1.7-8.0

Before you begin reading this guide, we recommend you run the Elasticsearch Error Check-Up which can resolve issues that cause many errors.

This guide will help help you understand why the log “Updating number_of_replicas to for indices” has appeared and how to resolve it. It’s also important to understand the issues related to the log, so you can read the general overview on common issues and tips related to these Elasticsearch concepts below: allocation, cluster, indices and routing.

<

Overview

Elasticsearch replica shards can be updated dynamically both manually (by hitting API) or dynamically (based on the number of nodes and configuration) in order to improve search latency. A change in the number of replicas is a critical event, and Elasticsearch logs this event as below:

updating number_of_replicas to {new replicas number} for indices {affected indices list}.

Manual update 

Using Index update setting for existing indices or Index template for a future index. You can update the replica count of an existing index using the update index setting API as shown below. 

PUT /my-index/_settings
{
  "index" : {
    "number_of_replicas" : 2 // replica count
  }
}

Dynamic update 

It’s common to upscale and downscale Elasticsearch clusters based on traffic, and for the same reason, having a fixed number of replicas isn’t very useful in cases like these. That’s why Elasticsearch came up with the concept of index.auto_expand_replicas(default false), which can dynamically increase/decrease the number of replicas based on the configuration and number of nodes in the cluster. 

Example:

If configured (0-5), the number of allocated replicas will match the number of nodes you create. If you increase/decrease, these numbers will also be changed accordingly.

Bottom line: 

This log is solely for information purposes and there’s nothing much to worry about with them, but if it changes without your knowledge, there is a scope to of appropriate actions to take, such as:

Changing the index template to make sure all future indices are not created with fixed settings,

Or:

index.auto_expand_replicas : false

Log Context

Log “Updating number_of_replicas to [{}] for indices {}” classname is AllocationService.java.
We extracted the following from Elasticsearch source code for those seeking an in-depth context :

                     final IndexMetaData indexMetaData = metaDataBuilder.get(index);
                    final IndexMetaData.Builder indexMetaDataBuilder =
                            new IndexMetaData.Builder(indexMetaData).settingsVersion(1 + indexMetaData.getSettingsVersion());
                    metaDataBuilder.put(indexMetaDataBuilder);
                }
                logger.info("updating number_of_replicas to [{}] for indices {}"; numberOfReplicas; indices);
            }
            final ClusterState fixedState = ClusterState.builder(clusterState).routingTable(routingTableBuilder.build())
                .metaData(metaDataBuilder).build();
            assert AutoExpandReplicas.getAutoExpandReplicaChanges(fixedState.metaData(); fixedState.nodes()).isEmpty();
            return fixedState;




 

Run the Check-Up to get customized insights on your system:

Overview

In Elasticsearch, an index (plural: indices) contains a schema and can have one or more shards and replicas. An Elasticsearch index is divided into shards and each shard is an instance of a Lucene index.

Indices are used to store the documents in dedicated data structures corresponding to the data type of fields. For example, text fields are stored inside an inverted index whereas numeric and geo fields are stored inside BKD trees.

Examples

Create index

The following example is based on Elasticsearch version 5.x onwards. An index with two shards, each having one replica will be created with the name test_index1

PUT /test_index1?pretty
{
    "settings" : {
        "number_of_shards" : 2,
        "number_of_replicas" : 1
    },
    "mappings" : {
        "properties" : {
            "tags" : { "type" : "keyword" },
            "updated_at" : { "type" : "date" }
        }
    }
}

List indices

All the index names and their basic information can be retrieved using the following command:

GET _cat/indices?v

Index a document

Let’s add a document in the index with the command below:

PUT test_index1/_doc/1
{
  "tags": [
    "opster",
    "elasticsearch"
  ],
  "date": "01-01-2020"
}

Query an index

GET test_index1/_search
{
  "query": {
    "match_all": {}
  }
}

Query multiple indices

It is possible to search multiple indices with a single request. If it is a raw HTTP request, index names should be sent in comma-separated format, as shown in the example below, and in the case of a query via a programming language client such as python or Java, index names are to be sent in a list format.

GET test_index1,test_index2/_search

Delete indices

DELETE test_index1

Common problems

  • It is good practice to define the settings and mapping of an Index wherever possible because if this is not done, Elasticsearch tries to automatically guess the data type of fields at the time of indexing. This automatic process may have disadvantages, such as mapping conflicts, duplicate data and incorrect data types being set in the index. If the fields are not known in advance, it’s better to use dynamic index templates.
  • Elasticsearch supports wildcard patterns in Index names, which sometimes aids with querying multiple indices, but can also be very destructive too. For example, It is possible to delete all the indices in a single command using the following commands:
DELETE /*

To disable this, you can add the following lines in the elasticsearch.yml:

action.destructive_requires_name: true

Log Context

Log “Updating number_of_replicas to [{}] for indices {}” classname is AllocationService.java.
We extracted the following from Elasticsearch source code for those seeking an in-depth context :

                     final IndexMetaData indexMetaData = metaDataBuilder.get(index);
                    final IndexMetaData.Builder indexMetaDataBuilder =
                            new IndexMetaData.Builder(indexMetaData).settingsVersion(1 + indexMetaData.getSettingsVersion());
                    metaDataBuilder.put(indexMetaDataBuilder);
                }
                logger.info("updating number_of_replicas to [{}] for indices {}"; numberOfReplicas; indices);
            }
            final ClusterState fixedState = ClusterState.builder(clusterState).routingTable(routingTableBuilder.build())
                .metaData(metaDataBuilder).build();
            assert AutoExpandReplicas.getAutoExpandReplicaChanges(fixedState.metaData(); fixedState.nodes()).isEmpty();
            return fixedState;




 

Run the Check-Up to get customized insights on your system:

Analyze your cluster

Skip to content