Package org.hibernate.search.backend.lucene.cfg

Class LuceneIndexSettings

java.lang.Object
org.hibernate.search.backend.lucene.cfg.LuceneIndexSettings
public final class LuceneIndexSettings extends Object
Configuration properties for Lucene indexes.

Constants in this class are to be appended to a prefix to form a property key; see IndexSettings for details.

  • Field Details

    • DIRECTORY_PREFIX

      public static final String DIRECTORY_PREFIX
      The prefix for directory-related property keys.
      See Also:

    • DIRECTORY_TYPE

      public static final String DIRECTORY_TYPE
      The type of directory to use when reading from or writing to the index.

      Expects a String, such as "local-filesystem". See the reference documentation for a list of available values.

      Defaults to LuceneIndexSettings.Defaults.DIRECTORY_TYPE.

      See Also:

    • DIRECTORY_ROOT

      public static final String DIRECTORY_ROOT
      The filesystem root for the directory.

      Only available for the "local-filesystem" directory type.

      Expects a String representing a path to an existing directory accessible in read and write mode, such as "local-filesystem".

      The actual index files will be created in directory <root>/<index-name>.

      Defaults to the JVM's working directory.

      See Also:

    • DIRECTORY_LOCKING_STRATEGY

      public static final String DIRECTORY_LOCKING_STRATEGY
      How to lock on the directory.

      Expects a LockingStrategyName value, or a String representation of such value.

      Defaults are specific to each directory type.

      See Also:

    • DIRECTORY_FILESYSTEM_ACCESS_STRATEGY

      public static final String DIRECTORY_FILESYSTEM_ACCESS_STRATEGY
      How to access the filesystem in the directory.

      Only available for the "local-filesystem" directory type.

      Expects a FileSystemAccessStrategyName value, or a String representation of such value.

      Defaults to LuceneIndexSettings.Defaults.DIRECTORY_FILESYSTEM_ACCESS_STRATEGY.

      See Also:

    • IO_PREFIX

      public static final String IO_PREFIX
      The prefix for I/O-related property keys.
      See Also:

    • IO_STRATEGY

      public static final String IO_STRATEGY
      How to handle input/output, i.e. how to write to and read from indexes.

      Expects a IOStrategyName value, or a String representation of such value.

      Defaults to LuceneIndexSettings.Defaults.IO_STRATEGY.

      See Also:

    • IO_COMMIT_INTERVAL

      public static final String IO_COMMIT_INTERVAL
      How much time may pass after an index change until the change is committed.

      Only available for the "near-real-time" I/O strategy.

      This effectively defines how long changes may be in an "unsafe" state, where a crash or power loss will result in data loss. For example:

      • if set to 0, changes are safe as soon as the background process finishes treating a batch of changes.
      • if set to 1000, changes may not be safe for an additional 1 second after the background process finishes treating a batch. There is a benefit, though: index changes occurring less than 1 second after another change may execute faster.

      Note that individual write operations may trigger a forced commit (for example with the write-sync and sync indexing plan synchronization strategies in the ORM mapper), in which case you will only benefit from a non-zero commit interval during intensive indexing (mass indexer, ...).

      Note that committing is not necessary to make changes visible to search queries: the two concepts are unrelated. See IO_REFRESH_INTERVAL.

      Expects a positive Integer value in milliseconds, such as 1000, or a String that can be parsed into such Integer value.

      Defaults to LuceneIndexSettings.Defaults.IO_COMMIT_INTERVAL.

      See Also:

    • IO_REFRESH_INTERVAL

      public static final String IO_REFRESH_INTERVAL
      How much time may pass after an index write until the index reader is considered stale and re-created.

      Only available for the "near-real-time" I/O strategy.

      This effectively defines how out-of-date search query results may be. For example:

      • If set to 0, search results will always be completely in sync with the index writes.
      • If set to 1000, search results may reflect the state of the index at most 1 second ago. There is a benefit, though: in situations where the index is being frequently written to, search queries executed less than 1 second after another query may execute faster.

      Note that individual write operations may trigger a forced refresh (for example with the read-sync and sync indexing plan synchronization strategies in the ORM mapper), in which case you will only benefit from a non-zero refresh interval during intensive indexing (mass indexer, ...).

      Expects a positive Integer value in milliseconds, such as 1000, or a String that can be parsed into such Integer value.

      Defaults to LuceneIndexSettings.Defaults.IO_REFRESH_INTERVAL.

      See Also:

    • IO_WRITER_PREFIX

      public static final String IO_WRITER_PREFIX
      The prefix for property keys related to the index writer.
      See Also:

    • IO_WRITER_MAX_BUFFERED_DOCS

      public static final String IO_WRITER_MAX_BUFFERED_DOCS
      The value to pass to IndexWriterConfig.setMaxBufferedDocs(int).

      Expects a positive Integer value, or a String that can be parsed into such Integer value.

      The default for this setting is defined by Lucene.

      See Also:

    • IO_WRITER_RAM_BUFFER_SIZE

      public static final String IO_WRITER_RAM_BUFFER_SIZE
      The value to pass to IndexWriterConfig.setRAMBufferSizeMB(double).

      Expects a positive Integer value in megabytes, or a String that can be parsed into such Integer value.

      The default for this setting is defined by Lucene.

      See Also:

    • IO_WRITER_INFOSTREAM

      public static final String IO_WRITER_INFOSTREAM
      Whether to log the IndexWriterConfig.setInfoStream(InfoStream) (at the trace level) or not.

      Logs are appended to the logger "org.hibernate.search.backend.lucene.infostream".

      Expects a Boolean value such as true or false, or a String that can be parsed into such Boolean value.

      Default is false.

      See Also:

    • IO_MERGE_PREFIX

      public static final String IO_MERGE_PREFIX
      The prefix for property keys related to merge.
      See Also:

    • IO_MERGE_MAX_DOCS

      public static final String IO_MERGE_MAX_DOCS
      The value to pass to LogMergePolicy.setMaxMergeDocs(int).

      Expects a positive Integer value, or a String that can be parsed into such Integer value.

      The default for this setting is defined by Lucene.

      See Also:

    • IO_MERGE_FACTOR

      public static final String IO_MERGE_FACTOR
      The value to pass to LogMergePolicy.setMergeFactor(int).

      Expects a positive Integer value, or a String that can be parsed into such Integer value.

      The default for this setting is defined by Lucene.

      See Also:

    • IO_MERGE_MIN_SIZE

      public static final String IO_MERGE_MIN_SIZE
      The value to pass to LogByteSizeMergePolicy.setMinMergeMB(double).

      Expects a positive Integer value in megabytes, or a String that can be parsed into such Integer value.

      The default for this setting is defined by Lucene.

      See Also:

    • IO_MERGE_MAX_SIZE

      public static final String IO_MERGE_MAX_SIZE
      The value to pass to LogByteSizeMergePolicy.setMaxMergeMB(double).

      Expects a positive Integer value in megabytes, or a String that can be parsed into such Integer value.

      The default for this setting is defined by Lucene.

      See Also:

    • IO_MERGE_MAX_FORCED_SIZE

      public static final String IO_MERGE_MAX_FORCED_SIZE
      The value to pass to LogByteSizeMergePolicy.setMaxMergeMBForForcedMerge(double).

      Expects a positive Integer value in megabytes, or a String that can be parsed into such Integer value.

      The default for this setting is defined by Lucene.

      See Also:

    • IO_MERGE_CALIBRATE_BY_DELETES

      public static final String IO_MERGE_CALIBRATE_BY_DELETES
      The value to pass to LogMergePolicy.setCalibrateSizeByDeletes(boolean).

      Expects a Boolean value such as true or false, or a String that can be parsed into such Boolean value.

      The default for this setting is defined by Lucene.

      See Also:

    • SHARDING_PREFIX

      public static final String SHARDING_PREFIX
      The prefix for sharding-related property keys.
      See Also:

    • SHARDING_STRATEGY

      public static final String SHARDING_STRATEGY
      The sharding strategy, deciding the number of shards, their identifiers, and how to translate a routing key into a shard identifier.

      Expects a String, such as "hash". See the reference documentation for a list of available values.

      Defaults to LuceneIndexSettings.Defaults.SHARDING_STRATEGY (no sharding).

      See Also:

    • SHARDING_NUMBER_OF_SHARDS

      public static final String SHARDING_NUMBER_OF_SHARDS
      The number of shards to create for the index, i.e. the number of "physical" indexes, each holding a part of the index data.

      Only available for the hash sharding strategy.

      Expects a strictly positive Integer value, such as 4, or a String that can be parsed into such Integer value.

      No default: this property must be set when using the hash sharding strategy.

      See Also:

    • SHARDING_SHARD_IDENTIFIERS

      public static final String SHARDING_SHARD_IDENTIFIERS
      The list of shard identifiers to accept for the index.

      Only available for the explicit sharding strategy.

      Expects either a String containing multiple shard identifiers separated by commas (','), or a Collection<String> containing such shard identifiers.

      No default: this property must be set when using the explicit sharding strategy.

      See Also:

    • SHARDS

      public static final String SHARDS
      The root property whose children are shards, e.g. shards.0.<some shard-scoped property> = bar or shards.1.<some shard-scoped property> = bar or shards.main.<some shard-scoped property> = bar.
      See Also:

    • INDEXING_PREFIX

      public static final String INDEXING_PREFIX
      The prefix for indexing-related property keys.
      See Also:

    • INDEXING_QUEUE_COUNT

      public static final String INDEXING_QUEUE_COUNT
      The number of indexing queues assigned to each index (or each shard of each index, when sharding is enabled).

      Expects a strictly positive integer value, or a string that can be parsed into an integer value.

      See the reference documentation, section "Lucene backend - Indexing", for more information about this setting and its implications.

      Defaults to LuceneIndexSettings.Defaults.INDEXING_QUEUE_COUNT.

      See Also:

    • INDEXING_QUEUE_SIZE

      public static final String INDEXING_QUEUE_SIZE
      The size of indexing queues.

      Expects a strictly positive integer value, or a string that can be parsed into an integer value.

      See the reference documentation, section "Lucene backend - Indexing", for more information about this setting and its implications.

      Defaults to LuceneIndexSettings.Defaults.INDEXING_QUEUE_SIZE.

      See Also:

  • Method Details

    • shardKey

      public static String shardKey(String shardId, String radical)
      Builds a configuration property key for the given shard of all indexes of the default backend, with the given radical.

      See constants in this class for available radicals.

      Example result: "hibernate.search.backend.shard.<shardId>.indexing.queue_count"
      Parameters:
      shardId - The identifier of the shard to configure.
      radical - The radical of the configuration property (see constants in this class).
      Returns:
      the concatenated shard settings key

    • shardKey

      public static String shardKey(String indexName, String shardId, String radical)
      Builds a configuration property key for the given shard of the given index of the default backend, with the given radical.

      See constants in this class for available radicals.

      Example result: "hibernate.search.backends.<backendName>.indexes.<indexName>.shard.<shardId>.indexing.queue_count"
      Parameters:
      indexName - The name of the index in which the shard to configure is located.
      shardId - The identifier of the shard to configure.
      radical - The radical of the configuration property (see constants in this class).
      Returns:
      the concatenated shard settings key

    • shardKey

      public static String shardKey(String backendName, String indexName, String shardId, String radical)
      Builds a configuration property key for the given shard of the given index of the given backend, with the given radical.

      See constants in this class for available radicals.

      Example result: "hibernate.search.backends.<backendName>.indexes.<indexName>.shard.<shardId>.indexing.queue_count"
      Parameters:
      backendName - The name of the backend in which the shard to configure is located.
      indexName - The name of the index in which the shard to configure is located.
      shardId - The identifier of the shard to configure.
      radical - The radical of the configuration property (see constants in this class).
      Returns:
      the concatenated shard settings key