Optimizing Apache Solr Schema and Configuration for Performance

1. Configure Caches

Solr caches are tied to a specific IndexSearcher instance; configuring them is crucial for performance.

Configure filterCache

The filterCache is used by SolrIndexSearcher for filter queries, allowing pre‑warming from the previous searcher. Example configuration:

<filterCache class="solr.FastLRUCache" size="512" initialSize="512" autowarmCount="0" />

class: SolrCache implementation (LRUCache or FastLRUCache) size: maximum entries initialSize: initial capacity autowarmCount: entries to pre‑warm from old cache

Configure queryResultCache and documentCache

queryResultCache stores ordered DocList of document IDs for previous searches; documentCache stores Lucene Document objects. They are useful for read‑heavy use cases such as a blog with many reads.

<queryResultCache class="solr.LRUCache" size="512" initialSize="512" autowarmCount="0" />
<documentCache class="solr.LRUCache" size="512" initialSize="512" autowarmCount="0" />

For write‑heavy workloads, disable these caches during soft commits.

2. Configure SolrCloud

SolrCloud enables fault‑tolerant, highly available clusters with master‑slave replication. In the master solrconfig.xml, include the following:

<str name="confFiles">solrconfig_slave.xml:solrconfig.xml,x.xml,y.xml</str>

Refer to Solr documentation for details.

3. Configure Commits

Commit controls when indexed data becomes searchable. Hard commit (commit=true) writes all Lucene files to stable storage; soft commit (softCommit=true) makes changes visible quickly without durable storage, enabling near‑real‑time search.

AutoCommit

autoCommit can be set to trigger based on maxDocs or maxTime. Example:

<autoCommit>
  <maxDocs>20000</maxDocs>
  <maxTime>50000</maxTime>
  <openSearcher>false</openSearcher>
</autoCommit>

Disabling autoCommit during bulk imports can improve migration speed.

4. Configure Dynamic Fields

Dynamic fields allow wildcard field names, reducing the need to define every field explicitly. Example:

<dynamicField name="*.fieldname" type="boolean" multiValued="true" stored="true" />

Use them judiciously, as each unique field name consumes Lucene memory.

5. Configure Indexed vs Stored Fields

Set indexed="true" for fields that need to be searchable or faceted; set indexed="false" for fields only needed for retrieval. Example:

<field name="foo" type="int" stored="true" indexed="false"/>

This can reduce re‑indexing time.

6. Configure copyField

copyField copies values from source fields to a destination field, often used to create a catch‑all search field. Example:

<copyField source="*_abcd" dest="wxyz"/>

7. Use Filter Queries (fq)

fq limits the result set without affecting scoring and is cached independently. Example curl request:

POST
{
 "form_params": {
   "fq": "id=1234",
   "fl": "abc cde",
   "wt": "json"
 },
 "query": {
   "q": "*:*"
 }
}

8. Use Faceting

Faceting groups results by field values for aggregation. Example request:

{
 "form_params": {
   "fq": "fieldName:value",
   "fl": "fieldName",
   "facet": "true",
   "facet.mincount": 1,
   "facet.limit": -1,
   "facet.field": "fieldName",
   "wt": "json"
 },
 "query": {
   "q": "*:*"
 }
}

Conclusion

When deploying Solr to production, tuning caches, commit settings, schema definitions, and using filter queries and faceting are essential for optimal performance, depending on the specific application workload.