[Java] Enable filluniquerandom, readseq, BloomFilter, and 70+ command-line options to DbBenchmark.java

Summary:
* Add filluniquerandom
* Add readseq, implemented using iterator.
* Realize most command-line-arguments from db_bench.cc (70+).
* Some code are commented out as some of the options in Options
  not yet have Java bindings.
* Add default option to DbBenchmark.
* RocksDB will now take the ownership of all c++ raw-pointers from Options, which includes a c++ raw-pointer for Filter.

Test Plan: ./jdb_bench.sh --db=/tmp/rocksjava-bench/db --num_levels=6 --key_size=20 --prefix_size=20 --keys_per_prefix=0 --value_size=100 --block_size=4096 --cache_size=17179869184 --cache_numshardbits=6 --compression_ratio=1 --min_level_to_compress=-1 --disable_seek_compaction=1 --hard_rate_limit=2 --write_buffer_size=134217728 --max_write_buffer_number=2 --level0_file_num_compaction_trigger=8 --target_file_size_base=134217728 --max_bytes_for_level_base=1073741824 --disable_wal=0 --wal_dir=/tmp/rocksjava-bench/wal --sync=0 --disable_data_sync=1 --verify_checksum=1 --delete_obsolete_files_period_micros=314572800 --max_grandparent_overlap_factor=10 --max_background_compactions=4 --max_background_flushes=0 --level0_slowdown_writes_trigger=16 --level0_stop_writes_trigger=24 --statistics=0 --stats_per_interval=0 --stats_interval=1048576 --histogram=0 --use_plain_table=1 --open_files=-1 --mmap_read=1 --mmap_write=0 --memtablerep=prefix_hash --bloom_bits=10 --bloom_locality=1 --perf_level=0 --benchmarks=filluniquerandom,readseq,readrandom --use_existing_db=0 --threads=4

Reviewers: haobo, dhruba, sdong, ankgup87, rsumbaly, swapnilghike, zzbennett

Reviewed By: haobo

CC: leveldb

Differential Revision: https://reviews.facebook.net/D18267

Author: yhchiang

java/org/rocksdb/Filter.java

@@ -20,7 +20,7 @@ public abstract class Filter {
   /**
    * Deletes underlying C++ filter pointer.
    */
-  public synchronized void dispose() {
+  protected synchronized void dispose() {
     if(nativeHandle_ != 0) {
       dispose0(nativeHandle_);
     }

java/org/rocksdb/Options.java

@@ -146,15 +146,21 @@ public long blockSize() {
 
   /**
    * Use the specified filter policy to reduce disk reads.
+   *
+   * Note that the caller should not dispose the input filter as
+   * Options.dispose() will dispose this filter.
+   *
    * @param Filter policy java instance.
    * @return the instance of the current Options.
    * @see RocksDB.open()
    */
   public Options setFilter(Filter filter) {
     assert(isInitialized());
-    setFilter0(nativeHandle_, filter);
+    setFilterHandle(nativeHandle_, filter.nativeHandle_);
+    filter_ = filter;
     return this;
   }
+  private native void setFilterHandle(long optHandle, long filterHandle);
 
   /*
    * Disable compaction triggered by seek.
@@ -786,7 +792,8 @@ private native void setTableCacheRemoveScanCountLimit(
       long handle, int limit);
 
   /**
-   * The following two fields affect how archived logs will be deleted.
+   * WalTtlSeconds() and walSizeLimitMB() affect how archived logs
+   * will be deleted.
    * 1. If both set to 0, logs will be deleted asap and will not get into
    *    the archive.
    * 2. If WAL_ttl_seconds is 0 and WAL_size_limit_MB is not 0,
@@ -800,6 +807,7 @@ private native void setTableCacheRemoveScanCountLimit(
    *    checks will be performed with ttl being first.
    *
    * @return the wal-ttl seconds
+   * @see walSizeLimitMB()
    */
   public long walTtlSeconds() {
     assert(isInitialized());
@@ -808,7 +816,8 @@ public long walTtlSeconds() {
   private native long walTtlSeconds(long handle);
 
   /**
-   * The following two fields affect how archived logs will be deleted.
+   * WalTtlSeconds() and walSizeLimitMB() affect how archived logs
+   * will be deleted.
    * 1. If both set to 0, logs will be deleted asap and will not get into
    *    the archive.
    * 2. If WAL_ttl_seconds is 0 and WAL_size_limit_MB is not 0,
@@ -823,13 +832,64 @@ public long walTtlSeconds() {
    *
    * @param walTtlSeconds the ttl seconds
    * @return the reference to the current option.
+   * @see setWalSizeLimitMB()
    */
-  public Options setWALTtlSeconds(long walTtlSeconds) {
+  public Options setWalTtlSeconds(long walTtlSeconds) {
     assert(isInitialized());
-    setWALTtlSeconds(nativeHandle_, walTtlSeconds);
+    setWalTtlSeconds(nativeHandle_, walTtlSeconds);
     return this;
   }
-  private native void setWALTtlSeconds(long handle, long walTtlSeconds);
+  private native void setWalTtlSeconds(long handle, long walTtlSeconds);
+
+  /**
+   * WalTtlSeconds() and walSizeLimitMB() affect how archived logs
+   * will be deleted.
+   * 1. If both set to 0, logs will be deleted asap and will not get into
+   *    the archive.
+   * 2. If WAL_ttl_seconds is 0 and WAL_size_limit_MB is not 0,
+   *    WAL files will be checked every 10 min and if total size is greater
+   *    then WAL_size_limit_MB, they will be deleted starting with the
+   *    earliest until size_limit is met. All empty files will be deleted.
+   * 3. If WAL_ttl_seconds is not 0 and WAL_size_limit_MB is 0, then
+   *    WAL files will be checked every WAL_ttl_secondsi / 2 and those that
+   *    are older than WAL_ttl_seconds will be deleted.
+   * 4. If both are not 0, WAL files will be checked every 10 min and both
+   *    checks will be performed with ttl being first.
+   *
+   * @return size limit in mega-bytes.
+   * @see walSizeLimitMB()
+   */
+  public long walSizeLimitMB() {
+    assert(isInitialized());
+    return walSizeLimitMB(nativeHandle_);
+  }
+  private native long walSizeLimitMB(long handle);
+
+  /**
+   * WalTtlSeconds() and walSizeLimitMB() affect how archived logs
+   * will be deleted.
+   * 1. If both set to 0, logs will be deleted asap and will not get into
+   *    the archive.
+   * 2. If WAL_ttl_seconds is 0 and WAL_size_limit_MB is not 0,
+   *    WAL files will be checked every 10 min and if total size is greater
+   *    then WAL_size_limit_MB, they will be deleted starting with the
+   *    earliest until size_limit is met. All empty files will be deleted.
+   * 3. If WAL_ttl_seconds is not 0 and WAL_size_limit_MB is 0, then
+   *    WAL files will be checked every WAL_ttl_secondsi / 2 and those that
+   *    are older than WAL_ttl_seconds will be deleted.
+   * 4. If both are not 0, WAL files will be checked every 10 min and both
+   *    checks will be performed with ttl being first.
+   *
+   * @param sizeLimitMB size limit in mega-bytes.
+   * @return the reference to the current option.
+   * @see setWalSizeLimitMB()
+   */
+  public Options setWalSizeLimitMB(long sizeLimitMB) {
+    assert(isInitialized());
+    setWalSizeLimitMB(nativeHandle_, sizeLimitMB);
+    return this;
+  }
+  private native void setWalSizeLimitMB(long handle, long sizeLimitMB);
 
   /**
    * Number of bytes to preallocate (via fallocate) the manifest
@@ -2298,8 +2358,7 @@ private native void setMaxBackgroundCompactions(
   private native void useFixedLengthPrefixExtractor(
       long handle, int prefixLength);
 
-  private native void setFilter0(long optHandle, Filter fp);
-
   long nativeHandle_;
   long cacheSize_;
+  Filter filter_;
 }

java/org/rocksdb/RocksDB.java

@@ -36,6 +36,7 @@ public static RocksDB open(String path) throws RocksDBException {
     // the c++ one.
     Options options = new Options();
     db.open(options.nativeHandle_, options.cacheSize_, path);
+    db.transferCppRawPointersOwnership(options);
     options.dispose();
     return db;
   }
@@ -46,8 +47,12 @@ public static RocksDB open(String path) throws RocksDBException {
    */
   public static RocksDB open(Options options, String path)
       throws RocksDBException {
+    // when non-default Options is used, keeping an Options reference
+    // in RocksDB can prevent Java to GC during the life-time of
+    // the currently-created RocksDB.
     RocksDB db = new RocksDB();
     db.open(options.nativeHandle_, options.cacheSize_, path);
+    db.transferCppRawPointersOwnershipFrom(options);
     return db;
   }
 
@@ -195,6 +200,17 @@ protected RocksDB() {
     nativeHandle_ = 0;
   }
 
+  /**
+   * Transfer the ownership of all c++ raw-pointers from Options
+   * to RocksDB to ensure the life-time of those raw-pointers
+   * will be at least as long as the life-time of any RocksDB
+   * that uses these raw-pointers.
+   */
+  protected void transferCppRawPointersOwnershipFrom(Options opt) {
+    filter_ = opt.filter_;
+    opt.filter_ = null;
+  }
+
   // native methods
   protected native void open(
       long optionsHandle, long cacheSize, String path) throws RocksDBException;
@@ -227,4 +243,5 @@ protected native void remove(
   protected native void close0();
 
   protected long nativeHandle_;
+  protected Filter filter_;
 }