com.bigdata.resources
Interface StoreManager.Options

All Superinterfaces:

Options

All Known Subinterfaces:

DataServer.Options, DataService.Options, IndexManager.Options, MetadataServer.Options, MetadataService.Options, OverflowManager.Options, ResourceManager.Options

Enclosing class:

StoreManager

public static interface StoreManager.Options
extends Options

Options for the StoreManager.

Version:

$Id: StoreManager.java 6271 2012-04-11 14:34:58Z thompsonbry $

Author:

Bryan Thompson

Field Summary

static String	ACCELERATE_OVERFLOW_THRESHOLD Option specifies the #of bytes under management below which we will accelerate the overflow of the live journal by reducing its maximum extent below the nominal configured maximum extent.
static String	DATA_DIR The property whose value is the name of the directory in which the store files will be created (no default).
static String	DEFAULT_ACCELERATE_OVERFLOW_THRESHOLD
static String	DEFAULT_IGNORE_BAD_FILES
static String	DEFAULT_PURGE_OLD_RESOURCES_DURING_STARTUP
static String	DEFAULT_STORE_CACHE_CAPACITY The default for the STORE_CACHE_CAPACITY option.
static String	DEFAULT_STORE_CACHE_TIMEOUT
static String	IGNORE_BAD_FILES A boolean property whose value determines whether or not startup will complete successfully if bad files are identified during the startup scan (default "false").
static String	PURGE_OLD_RESOURCES_DURING_STARTUP Option may be used to disable the purge of old resources during startup.
static String	STORE_CACHE_CAPACITY The capacity of the LRU cache of open IRawStores.
static String	STORE_CACHE_TIMEOUT The time in milliseconds before an entry in the store cache will be cleared from the backing HardReferenceQueue (default "60000").

Fields inherited from interface com.bigdata.journal.Options

ALTERNATE_ROOT_BLOCK, BUFFER_MODE, CREATE, CREATE_TEMP_FILE, CREATE_TIME, DEFAULT_BUFFER_MODE, DEFAULT_CREATE, DEFAULT_CREATE_TEMP_FILE, DEFAULT_DELETE_ON_CLOSE, DEFAULT_DELETE_ON_EXIT, DEFAULT_DOUBLE_SYNC, DEFAULT_FILE_LOCK_ENABLED, DEFAULT_FORCE_ON_COMMIT, DEFAULT_FORCE_WRITES, DEFAULT_HISTORICAL_INDEX_CACHE_CAPACITY, DEFAULT_HISTORICAL_INDEX_CACHE_TIMEOUT, DEFAULT_INITIAL_EXTENT, DEFAULT_LIVE_INDEX_CACHE_CAPACITY, DEFAULT_LIVE_INDEX_CACHE_TIMEOUT, DEFAULT_MAXIMUM_EXTENT, DEFAULT_MINIMUM_EXTENSION, DEFAULT_READ_CACHE_CAPACITY, DEFAULT_READ_CACHE_MAX_RECORD_SIZE, DEFAULT_READ_ONLY, DEFAULT_USE_DIRECT_BUFFERS, DEFAULT_VALIDATE_CHECKSUM, DEFAULT_WRITE_CACHE_BUFFER_COUNT, DEFAULT_WRITE_CACHE_ENABLED, DELETE_ON_CLOSE, DELETE_ON_EXIT, DOUBLE_SYNC, FILE, FILE_LOCK_ENABLED, FORCE_ON_COMMIT, FORCE_WRITES, HISTORICAL_INDEX_CACHE_CAPACITY, HISTORICAL_INDEX_CACHE_TIMEOUT, IGNORE_BAD_ROOT_BLOCK, INITIAL_EXTENT, JNL, LIVE_INDEX_CACHE_CAPACITY, LIVE_INDEX_CACHE_TIMEOUT, MAXIMUM_EXTENT, MEM_MAX_EXTENT, MINIMUM_EXTENSION, minimumInitialExtent, minimumMinimumExtension, OFFSET_BITS, OTHER_MAX_EXTENT, READ_ONLY, RW_MAX_EXTENT, SEG, TMP_DIR, UPDATE_ICU_VERSION, USE_DIRECT_BUFFERS, VALIDATE_CHECKSUM, WRITE_CACHE_BUFFER_COUNT, WRITE_CACHE_ENABLED

Field Detail

DATA_DIR

static final String DATA_DIR

The property whose value is the name of the directory in which the store files will be created (no default). This property is required unless the instance is transient. If you specify Options.BUFFER_MODE as BufferMode.Transient then journals will be NOT stored in the file system and OverflowManager.overflow() will be disabled.

The files are created within subdirectories as follows: The "journals" subdirectory contains the journal files. The "segments" directory contains subdirectories corresponding to the index UUID for each scale-out index. Within those index-specific directories, the index segment files are assigned to files using the temporary file mechanisms using the munged index name as the file prefix and Options.SEG as the file suffix. If the index is partitioned then the partition identifier appears as part of the file prefix.

Note: While files are stored per the scheme described above, the entire DATA_DIR will be scanned recursively to identify all journal files and index segments during startup. Files will be used where ever they are found but the IResourceMetadata.getFile() read from a given resource MUST correspond to its relative location within the DATA_DIR.

Note: Each DataService or MetadataService MUST have its own DATA_DIR.

STORE_CACHE_CAPACITY

static final String STORE_CACHE_CAPACITY

The capacity of the LRU cache of open IRawStores. The capacity of this cache indirectly controls how many stores will be held open. The main reason for keeping an store open is to reuse its buffers if another request arrives "soon" which would read on that store. Note that "stores" includes both StoreManager.ManagedJournals and IndexSegmentStores.

The effect of this parameter is indirect owning to the semantics of weak references and the control of the JVM over when they are cleared. Once an index becomes weakly reachable, the JVM will eventually GC the index object, thereby releasing its object graph. Since stores which are strongly reachable never have their weak reference cleared this provides our guarantee that stores are never closed if they are in use.

Stores have non-transient resources and MUST explicitly be closed. Since we are not notified before the weak reference is closed, our only remaining option is AbstractJournal.finalize() and IndexSegmentStore.finalize(), both of which close the store if it is still open.

See Also:

DEFAULT_STORE_CACHE_CAPACITY

DEFAULT_STORE_CACHE_CAPACITY

static final String DEFAULT_STORE_CACHE_CAPACITY

The default for the STORE_CACHE_CAPACITY option.

See Also:

Constant Field Values

STORE_CACHE_TIMEOUT

static final String STORE_CACHE_TIMEOUT

The time in milliseconds before an entry in the store cache will be cleared from the backing HardReferenceQueue (default "60000"). This property controls how long the store cache will retain an IRawStore which has not been recently used. This is in contrast to the cache capacity.

DEFAULT_STORE_CACHE_TIMEOUT

static final String DEFAULT_STORE_CACHE_TIMEOUT

See Also:

Constant Field Values

IGNORE_BAD_FILES

static final String IGNORE_BAD_FILES

A boolean property whose value determines whether or not startup will complete successfully if bad files are identified during the startup scan (default "false"). When false the StoreManager will refuse to start if if find bad files. When true the StoreManager will startup anyway but some index views may not be available. Regardless, bad files will be logged as they are identified and all files will be scanned before the StoreManager aborts.

DEFAULT_IGNORE_BAD_FILES

static final String DEFAULT_IGNORE_BAD_FILES

See Also:

Constant Field Values

PURGE_OLD_RESOURCES_DURING_STARTUP

static final String PURGE_OLD_RESOURCES_DURING_STARTUP

Option may be used to disable the purge of old resources during startup.

DEFAULT_PURGE_OLD_RESOURCES_DURING_STARTUP

static final String DEFAULT_PURGE_OLD_RESOURCES_DURING_STARTUP

See Also:

Constant Field Values

ACCELERATE_OVERFLOW_THRESHOLD

static final String ACCELERATE_OVERFLOW_THRESHOLD

Option specifies the #of bytes under management below which we will accelerate the overflow of the live journal by reducing its maximum extent below the nominal configured maximum extent. The purpose of this option is to promote rapid overflow of a new data service (where new is measured by the #of bytes under management). This helps to increase the rate at which index partitions are split (and moved if the there is more than one new data service starting). When ZERO (0) the feature is disabled.

DEFAULT_ACCELERATE_OVERFLOW_THRESHOLD

static final String DEFAULT_ACCELERATE_OVERFLOW_THRESHOLD

See Also:

Constant Field Values