Briefly, this error occurs when Elasticsearch attempts to add a legacy index template because it doesn’t exist in the system. This could be due to a missing or incorrectly configured template. To resolve this issue, you can either create the missing template or correct the existing one. Ensure that the template is properly configured according to the Elasticsearch version you are using. Also, check your Elasticsearch logs for more details about the missing template. Lastly, consider upgrading to the latest version of Elasticsearch, as legacy templates are deprecated in newer versions.
This guide will help you check for common problems that cause the log ” adding legacy template [{}] for [{}]; because it doesn’t exist ” to appear. To understand the issues related to this log, read the explanation below about the following Elasticsearch concepts: template, plugin.
If you want to learn more about Elasticsearch templates, check out this guide.
Overview
A template in Elasticsearch falls into one of the two following categories and is indexed inside Elasticsearch using its dedicated endpoint:
- Index templates, which are a way to define a set of rules including index settings, mappings and an index pattern. The template is applied automatically whenever a new index is created with the matching pattern. Templates are also used to dynamically apply custom mapping for the fields which are not predefined inside existing mapping.
- Search templates, which help in defining templates for search queries using mustache scripting language. These templates act as a placeholder for variables defined inside the search queries.
Examples
Create a dynamic index template
PUT /_template/template_1?pretty
{
"index_patterns": [
"logs*",
"api*"
],
"settings": {
"number_of_shards": 2
},
"mappings": {
"dynamic_templates": [
{
"strings": {
"match_mapping_type": "string",
"mapping": {
"type": "keyword"
}
}
}
],
"properties": {
"host_name": {
"type": "keyword"
},
"created_at": {
"type": "date"
}
}
}
}Create a search template
POST /_scripts/search_template_1?pretty
{
"script": {
"lang": "mustache",
"source": {
"query": {
"match": {
"description": "{{query_string}}"
}
}
}
}
}Executing a search query using search template
GET /_search/template?pretty
{
"id": "search_template_1",
"params": {
"query_string": "hello world"
}
}The search request will be executed by default on all the indices available in the cluster and can be limited to particular indices using an index parameter.
Notes
- A dynamic index template is always useful when you do not know the field names in advance and want to control their mapping as per the business use case.
Log Context
Log “adding legacy template [{}] for [{}]; because it doesn’t exist” classname is IndexTemplateRegistry.java.
We extracted the following from Elasticsearch source code for those seeking an in-depth context :
final String templateName = newTemplate.getTemplateName();
final AtomicBoolean creationCheck = templateCreationsInProgress.computeIfAbsent(templateName; key -> new AtomicBoolean(false));
if (creationCheck.compareAndSet(false; true)) {
IndexTemplateMetadata currentTemplate = state.metadata().getTemplates().get(templateName);
if (Objects.isNull(currentTemplate)) {
logger.info("adding legacy template [{}] for [{}]; because it doesn't exist"; templateName; getOrigin());
putLegacyTemplate(newTemplate; creationCheck);
} else if (Objects.isNull(currentTemplate.getVersion()) || newTemplate.getVersion() > currentTemplate.getVersion()) {
// IndexTemplateConfig now enforces templates contain a `version` property; so if the template doesn't have one we can
// safely assume it's an old version of the template.
logger.info(
