rocksdb/doc/index.html

<!DOCTYPE html>
<html>
<head>
<link rel="stylesheet" type="text/css" href="doc.css" />
<title>RocksDB</title>
</head>

<body>
<h1>RocksDB</h1>
<address>The Facebook Database Engineering Team</address>
<address>Build on earlier work on leveldb by Sanjay Ghemawat
               (sanjay@google.com) and Jeff Dean (jeff@google.com)</address>
<p>
The <code>rocksdb</code> library provides a persistent key value store.  Keys and
values are arbitrary byte arrays.  The keys are ordered within the key
value store according to a user-specified comparator function.

<p>
<h1>Opening A Database</h1>
<p>
A <code>rocksdb</code> database has a name which corresponds to a file system
directory.  All of the contents of database are stored in this
directory.  The following example shows how to open a database,
creating it if necessary:
<p>
<pre>
  #include &lt;assert&gt;
  #include "rocksdb/db.h"

  rocksdb::DB* db;
  rocksdb::Options options;
  options.create_if_missing = true;
  rocksdb::Status status = rocksdb::DB::Open(options, "/tmp/testdb", &amp;db);
  assert(status.ok());
  ...
</pre>
If you want to raise an error if the database already exists, add
the following line before the <code>rocksdb::DB::Open</code> call:
<pre>
  options.error_if_exists = true;
</pre>
<h1>Status</h1>
<p>
You may have noticed the <code>rocksdb::Status</code> type above.  Values of this
type are returned by most functions in <code>rocksdb</code> that may encounter an
error.  You can check if such a result is ok, and also print an
associated error message:
<p>
<pre>
   rocksdb::Status s = ...;
   if (!s.ok()) cerr &lt;&lt; s.ToString() &lt;&lt; endl;
</pre>
<h1>Closing A Database</h1>
<p>
When you are done with a database, just delete the database object.
Example:
<p>
<pre>
  ... open the db as described above ...
  ... do something with db ...
  delete db;
</pre>
<h1>Reads And Writes</h1>
<p>
The database provides <code>Put</code>, <code>Delete</code>, and <code>Get</code> methods to
modify/query the database.  For example, the following code
moves the value stored under key1 to key2.
<pre>
  std::string value;
  rocksdb::Status s = db-&gt;Get(rocksdb::ReadOptions(), key1, &amp;value);
  if (s.ok()) s = db-&gt;Put(rocksdb::WriteOptions(), key2, value);
  if (s.ok()) s = db-&gt;Delete(rocksdb::WriteOptions(), key1);
</pre>

<h1>Atomic Updates</h1>
<p>
Note that if the process dies after the Put of key2 but before the
delete of key1, the same value may be left stored under multiple keys.
Such problems can be avoided by using the <code>WriteBatch</code> class to
atomically apply a set of updates:
<p>
<pre>
  #include "leveldb/write_batch.h"
  ...
  std::string value;
  rocksdb::Status s = db-&gt;Get(rocksdb::ReadOptions(), key1, &amp;value);
  if (s.ok()) {
    rocksdb::WriteBatch batch;
    batch.Delete(key1);
    batch.Put(key2, value);
    s = db-&gt;Write(rocksdb::WriteOptions(), &amp;batch);
  }
</pre>
The <code>WriteBatch</code> holds a sequence of edits to be made to the database,
and these edits within the batch are applied in order.  Note that we
called <code>Delete</code> before <code>Put</code> so that if <code>key1</code> is identical to <code>key2</code>,
we do not end up erroneously dropping the value entirely.
<p>
Apart from its atomicity benefits, <code>WriteBatch</code> may also be used to
speed up bulk updates by placing lots of individual mutations into the
same batch.

<h1>Synchronous Writes</h1>
By default, each write to <code>leveldb</code> is asynchronous: it
returns after pushing the write from the process into the operating
system.  The transfer from operating system memory to the underlying
persistent storage happens asynchronously.  The <code>sync</code> flag
can be turned on for a particular write to make the write operation
not return until the data being written has been pushed all the way to
persistent storage.  (On Posix systems, this is implemented by calling
either <code>fsync(...)</code> or <code>fdatasync(...)</code> or
<code>msync(..., MS_SYNC)</code> before the write operation returns.)
<pre>
  rocksdb::WriteOptions write_options;
  write_options.sync = true;
  db-&gt;Put(write_options, ...);
</pre>
Asynchronous writes are often more than a thousand times as fast as
synchronous writes.  The downside of asynchronous writes is that a
crash of the machine may cause the last few updates to be lost.  Note
that a crash of just the writing process (i.e., not a reboot) will not
cause any loss since even when <code>sync</code> is false, an update
is pushed from the process memory into the operating system before it
is considered done.

<p>
Asynchronous writes can often be used safely.  For example, when
loading a large amount of data into the database you can handle lost
updates by restarting the bulk load after a crash.  A hybrid scheme is
also possible where every Nth write is synchronous, and in the event
of a crash, the bulk load is restarted just after the last synchronous
write finished by the previous run.  (The synchronous write can update
a marker that describes where to restart on a crash.)

<p>
<code>WriteBatch</code> provides an alternative to asynchronous writes.
Multiple updates may be placed in the same <code>WriteBatch</code> and
applied together using a synchronous write (i.e.,
<code>write_options.sync</code> is set to true).  The extra cost of
the synchronous write will be amortized across all of the writes in
the batch.

<p>
We also provide a way to completely disable Write Ahead Log for a
particular write. If you set write_option.disableWAL to true, the
write will not go to the log at all and may be lost in an event of
process crash.

<p>
When opening a DB, you can disable syncing of data files by setting
Options::disableDataSync to true. This can be useful when doing
bulk-loading or big idempotent operations. Once the operation is
finished, you can manually call sync() to flush all dirty buffers
to stable storage.

<p>
RocksDB by default uses faster fdatasync() to sync files. If you want
to use fsync(), you can set Options::use_fsync to true. You should set
this to true on filesystems like ext3 that can lose files after a
reboot.

<p>
<h1>Concurrency</h1>
<p>
A database may only be opened by one process at a time.
The <code>rocksdb</code> implementation acquires a lock from the
operating system to prevent misuse.  Within a single process, the
same <code>rocksdb::DB</code> object may be safely shared by multiple
concurrent threads.  I.e., different threads may write into or fetch
iterators or call <code>Get</code> on the same database without any
external synchronization (the leveldb implementation will
automatically do the required synchronization).  However other objects
(like Iterator and WriteBatch) may require external synchronization.
If two threads share such an object, they must protect access to it
using their own locking protocol.  More details are available in
the public header files.

<p>
<h1>Merge operators</h1>
<p>
Merge operators provide efficient support for read-modify-write operation.
More on the interface and implementation can be found on:
<p>
<a href="https://github.com/facebook/rocksdb/wiki/Merge-Operator">
    Merge Operator</a>
<p>
<a href="https://github.com/facebook/rocksdb/wiki/Merge-Operator-Implementation">
    Merge Operator Implementation</a>

<p>
<h1>Iteration</h1>
<p>
The following example demonstrates how to print all key,value pairs
in a database.
<p>
<pre>
  rocksdb::Iterator* it = db-&gt;NewIterator(rocksdb::ReadOptions());
  for (it-&gt;SeekToFirst(); it-&gt;Valid(); it-&gt;Next()) {
    cout &lt;&lt; it-&gt;key().ToString() &lt;&lt; ": "  &lt;&lt; it-&gt;value().ToString() &lt;&lt; endl;
  }
  assert(it-&gt;status().ok());  // Check for any errors found during the scan
  delete it;
</pre>
The following variation shows how to process just the keys in the
range <code>[start,limit)</code>:
<p>
<pre>
  for (it-&gt;Seek(start);
       it-&gt;Valid() &amp;&amp; it-&gt;key().ToString() &lt; limit;
       it-&gt;Next()) {
    ...
  }
</pre>
You can also process entries in reverse order.  (Caveat: reverse
iteration may be somewhat slower than forward iteration.)
<p>
<pre>
  for (it-&gt;SeekToLast(); it-&gt;Valid(); it-&gt;Prev()) {
    ...
  }
</pre>
<h1>Snapshots</h1>
<p>
Snapshots provide consistent read-only views over the entire state of
the key-value store.  <code>ReadOptions::snapshot</code> may be non-NULL to indicate
that a read should operate on a particular version of the DB state.
If <code>ReadOptions::snapshot</code> is NULL, the read will operate on an
implicit snapshot of the current state.
<p>
Snapshots are created by the DB::GetSnapshot() method:
<p>
<pre>
  rocksdb::ReadOptions options;
  options.snapshot = db-&gt;GetSnapshot();
  ... apply some updates to db ...
  rocksdb::Iterator* iter = db-&gt;NewIterator(options);
  ... read using iter to view the state when the snapshot was created ...
  delete iter;
  db-&gt;ReleaseSnapshot(options.snapshot);
</pre>
Note that when a snapshot is no longer needed, it should be released
using the DB::ReleaseSnapshot interface.  This allows the
implementation to get rid of state that was being maintained just to
support reading as of that snapshot.
<h1>Slice</h1>
<p>
The return value of the <code>it->key()</code> and <code>it->value()</code> calls above
are instances of the <code>rocksdb::Slice</code> type.  <code>Slice</code> is a simple
structure that contains a length and a pointer to an external byte
array.  Returning a <code>Slice</code> is a cheaper alternative to returning a
<code>std::string</code> since we do not need to copy potentially large keys and
values.  In addition, <code>rocksdb</code> methods do not return null-terminated
C-style strings since <code>rocksdb</code> keys and values are allowed to
contain '\0' bytes.
<p>
C++ strings and null-terminated C-style strings can be easily converted
to a Slice:
<p>
<pre>
   rocksdb::Slice s1 = "hello";

   std::string str("world");
   rocksdb::Slice s2 = str;
</pre>
A Slice can be easily converted back to a C++ string:
<pre>
   std::string str = s1.ToString();
   assert(str == std::string("hello"));
</pre>
Be careful when using Slices since it is up to the caller to ensure that
the external byte array into which the Slice points remains live while
the Slice is in use.  For example, the following is buggy:
<p>
<pre>
   rocksdb::Slice slice;
   if (...) {
     std::string str = ...;
     slice = str;
   }
   Use(slice);
</pre>
When the <code>if</code> statement goes out of scope, <code>str</code> will be destroyed and the
backing storage for <code>slice</code> will disappear.
<p>
<h1>Comparators</h1>
<p>
The preceding examples used the default ordering function for key,
which orders bytes lexicographically.  You can however supply a custom
comparator when opening a database.  For example, suppose each
database key consists of two numbers and we should sort by the first
number, breaking ties by the second number.  First, define a proper
subclass of <code>rocksdb::Comparator</code> that expresses these rules:
<p>
<pre>
  class TwoPartComparator : public rocksdb::Comparator {
   public:
    // Three-way comparison function:
    //   if a &lt; b: negative result
    //   if a &gt; b: positive result
    //   else: zero result
    int Compare(const rocksdb::Slice&amp; a, const rocksdb::Slice&amp; b) const {
      int a1, a2, b1, b2;
      ParseKey(a, &amp;a1, &amp;a2);
      ParseKey(b, &amp;b1, &amp;b2);
      if (a1 &lt; b1) return -1;
      if (a1 &gt; b1) return +1;
      if (a2 &lt; b2) return -1;
      if (a2 &gt; b2) return +1;
      return 0;
    }

    // Ignore the following methods for now:
    const char* Name() const { return "TwoPartComparator"; }
    void FindShortestSeparator(std::string*, const rocksdb::Slice&amp;) const { }
    void FindShortSuccessor(std::string*) const { }
  };
</pre>
Now create a database using this custom comparator:
<p>
<pre>
  TwoPartComparator cmp;
  rocksdb::DB* db;
  rocksdb::Options options;
  options.create_if_missing = true;
  options.comparator = &amp;cmp;
  rocksdb::Status status = rocksdb::DB::Open(options, "/tmp/testdb", &amp;db);
  ...
</pre>
<h2>Backwards compatibility</h2>
<p>
The result of the comparator's <code>Name</code> method is attached to the
database when it is created, and is checked on every subsequent
database open.  If the name changes, the <code>rocksdb::DB::Open</code> call will
fail.  Therefore, change the name if and only if the new key format
and comparison function are incompatible with existing databases, and
it is ok to discard the contents of all existing databases.
<p>
You can however still gradually evolve your key format over time with
a little bit of pre-planning.  For example, you could store a version
number at the end of each key (one byte should suffice for most uses).
When you wish to switch to a new key format (e.g., adding an optional
third part to the keys processed by <code>TwoPartComparator</code>),
(a) keep the same comparator name (b) increment the version number
for new keys (c) change the comparator function so it uses the
version numbers found in the keys to decide how to interpret them.


<p>
<h1>MemTable and Table factories</h1>
<p>
By default, we keep the data in memory in skiplist memtable and the data
on disk in a table format described here:
<a href="https://github.com/facebook/rocksdb/wiki/Rocksdb-Table-Format">
    RocksDB Table Format</a>.
<p>
Since one of the goals of RocksDB is to have
different parts of the system easily pluggable, we support different
implementations of both memtable and table format. You can supply
your own memtable factory by setting <code>Options::memtable_factory</code>
and your own table factory by setting <code>Options::table_factory</code>.
For available memtable factories, please refer to
<code>rocksdb/memtablerep.h</code> and for table factores to
<code>rocksdb/table.h</code>. These features are both in active development
and please be wary of any API changes that might break your application
going forward.
<p>
You can also read more about memtables here:
<a href="https://github.com/facebook/rocksdb/wiki/Rocksdb-Architecture-Guide#memtables">
Memtables wiki
</a>

<p>
<h1>Performance</h1>
<p>
Performance can be tuned by changing the default values of the
types defined in <code>include/rocksdb/options.h</code>.

<p>
<h2>Block size</h2>
<p>
<code>rocksdb</code> groups adjacent keys together into the same block and such a
block is the unit of transfer to and from persistent storage.  The
default block size is approximately 4096 uncompressed bytes.
Applications that mostly do bulk scans over the contents of the
database may wish to increase this size.  Applications that do a lot
of point reads of small values may wish to switch to a smaller block
size if performance measurements indicate an improvement.  There isn't
much benefit in using blocks smaller than one kilobyte, or larger than
a few megabytes.  Also note that compression will be more effective
with larger block sizes. To change block size parameter, use
<code>Options::block_size</code>.
<p>
<h2>Write buffer</h2>
<p>
<code>Options::write_buffer_size</code> specifies the amount of data
to build up in memory before converting to a sorted on-disk file.
Larger values increase performance, especially during bulk loads.
Up to max_write_buffer_number write buffers may be held in memory
at the same time,
so you may wish to adjust this parameter to control memory usage.
Also, a larger write buffer will result in a longer recovery time
the next time the database is opened.
Related option is
<code>Options::max_write_buffer_number</code>, which is maximum number
of write buffers that are built up in memory. The default is 2, so that
when 1 write buffer is being flushed to storage, new writes can continue
to the other write buffer.
<code>Options::min_write_buffer_number_to_merge</code> is the minimum number
of write buffers that will be merged together before writing to storage.
If set to 1, then all write buffers are fushed to L0 as individual files and
this increases read amplification because a get request has to check in all
of these files. Also, an in-memory merge may result in writing lesser
data to storage if there are duplicate records in each of these
individual write buffers.  Default: 1
<p>
<h2>Compression</h2>
<p>
Each block is individually compressed before being written to
persistent storage.  Compression is on by default since the default
compression method is very fast, and is automatically disabled for
uncompressible data.  In rare cases, applications may want to disable
compression entirely, but should only do so if benchmarks show a
performance improvement:
<p>
<pre>
  rocksdb::Options options;
  options.compression = rocksdb::kNoCompression;
  ... rocksdb::DB::Open(options, name, ...) ....
</pre>
<h2>Cache</h2>
<p>
The contents of the database are stored in a set of files in the
filesystem and each file stores a sequence of compressed blocks.  If
<code>options.block_cache</code> is non-NULL, it is used to cache frequently
used uncompressed block contents. If <code>options.block_cache_compressed</code>
is non-NULL, it is used to cache frequently used compressed blocks. Compressed
cache is an alternative to OS cache, which also caches compressed blocks. If
compressed cache is used, the OS cache will be disabled automatically by setting
<code>options.allow_os_buffer</code> to false.
<p>
<pre>
  #include "rocksdb/cache.h"

  rocksdb::Options options;
  options.block_cache = rocksdb::NewLRUCache(100 * 1048576);  // 100MB uncompressed cache
  options.block_cache_compressed = rocksdb::NewLRUCache(100 * 1048576);  // 100MB compressed cache
  rocksdb::DB* db;
  rocksdb::DB::Open(options, name, &db);
  ... use the db ...
  delete db
  delete options.block_cache;
  delete options.block_cache_compressed;
</pre>
<p>
When performing a bulk read, the application may wish to disable
caching so that the data processed by the bulk read does not end up
displacing most of the cached contents.  A per-iterator option can be
used to achieve this:
<p>
<pre>
  rocksdb::ReadOptions options;
  options.fill_cache = false;
  rocksdb::Iterator* it = db-&gt;NewIterator(options);
  for (it-&gt;SeekToFirst(); it-&gt;Valid(); it-&gt;Next()) {
    ...
  }
</pre>
<p>
You can also disable block cache by setting <code>options.no_block_cache</code>
to true.
<h2>Key Layout</h2>
<p>
Note that the unit of disk transfer and caching is a block.  Adjacent
keys (according to the database sort order) will usually be placed in
the same block.  Therefore the application can improve its performance
by placing keys that are accessed together near each other and placing
infrequently used keys in a separate region of the key space.
<p>
For example, suppose we are implementing a simple file system on top
of <code>rocksdb</code>.  The types of entries we might wish to store are:
<p>
<pre>
   filename -&gt; permission-bits, length, list of file_block_ids
   file_block_id -&gt; data
</pre>
We might want to prefix <code>filename</code> keys with one letter (say '/') and the
<code>file_block_id</code> keys with a different letter (say '0') so that scans
over just the metadata do not force us to fetch and cache bulky file
contents.
<p>
<h2>Filters</h2>
<p>
Because of the way <code>rocksdb</code> data is organized on disk,
a single <code>Get()</code> call may involve multiple reads from disk.
The optional <code>FilterPolicy</code> mechanism can be used to reduce
the number of disk reads substantially.
<pre>
   rocksdb::Options options;
   options.filter_policy = NewBloomFilter(10);
   rocksdb::DB* db;
   rocksdb::DB::Open(options, "/tmp/testdb", &amp;db);
   ... use the database ...
   delete db;
   delete options.filter_policy;
</pre>
The preceding code associates a
<a href="http://en.wikipedia.org/wiki/Bloom_filter">Bloom filter</a>
based filtering policy with the database.  Bloom filter based
filtering relies on keeping some number of bits of data in memory per
key (in this case 10 bits per key since that is the argument we passed
to NewBloomFilter).  This filter will reduce the number of unnecessary
disk reads needed for <code>Get()</code> calls by a factor of
approximately a 100.  Increasing the bits per key will lead to a
larger reduction at the cost of more memory usage.  We recommend that
applications whose working set does not fit in memory and that do a
lot of random reads set a filter policy.
<p>
If you are using a custom comparator, you should ensure that the filter
policy you are using is compatible with your comparator.  For example,
consider a comparator that ignores trailing spaces when comparing keys.
<code>NewBloomFilter</code> must not be used with such a comparator.
Instead, the application should provide a custom filter policy that
also ignores trailing spaces.  For example:
<pre>
  class CustomFilterPolicy : public rocksdb::FilterPolicy {
   private:
    FilterPolicy* builtin_policy_;
   public:
    CustomFilterPolicy() : builtin_policy_(NewBloomFilter(10)) { }
    ~CustomFilterPolicy() { delete builtin_policy_; }

    const char* Name() const { return "IgnoreTrailingSpacesFilter"; }

    void CreateFilter(const Slice* keys, int n, std::string* dst) const {
      // Use builtin bloom filter code after removing trailing spaces
      std::vector&lt;Slice&gt; trimmed(n);
      for (int i = 0; i &lt; n; i++) {
        trimmed[i] = RemoveTrailingSpaces(keys[i]);
      }
      return builtin_policy_-&gt;CreateFilter(&amp;trimmed[i], n, dst);
    }

    bool KeyMayMatch(const Slice& key, const Slice& filter) const {
      // Use builtin bloom filter code after removing trailing spaces
      return builtin_policy_-&gt;KeyMayMatch(RemoveTrailingSpaces(key), filter);
    }
  };
</pre>
<p>
Advanced applications may provide a filter policy that does not use
a bloom filter but uses some other mechanism for summarizing a set
of keys.  See <code>rocksdb/filter_policy.h</code> for detail.
<p>
<h1>Checksums</h1>
<p>
<code>rocksdb</code> associates checksums with all data it stores in the file system.
There are two separate controls provided over how aggressively these
checksums are verified:
<p>
<ul>
<li> <code>ReadOptions::verify_checksums</code> may be set to true to force
  checksum verification of all data that is read from the file system on
  behalf of a particular read.  By default, no such verification is
  done.
<p>
<li> <code>Options::paranoid_checks</code> may be set to true before opening a
  database to make the database implementation raise an error as soon as
  it detects an internal corruption.  Depending on which portion of the
  database has been corrupted, the error may be raised when the database
  is opened, or later by another database operation.  By default,
  paranoid checking is off so that the database can be used even if
  parts of its persistent storage have been corrupted.
<p>
  If a database is corrupted (perhaps it cannot be opened when
  paranoid checking is turned on), the <code>rocksdb::RepairDB</code> function
  may be used to recover as much of the data as possible.
<p>
</ul>

<p>
<h1>Compaction</h1>
<p>
You can read more on Compactions here:
<a href="https://github.com/facebook/rocksdb/wiki/Rocksdb-Architecture-Guide#multi-threaded-compactions">
    Multi-threaded compactions
</a>
<p>
Here we give overview of the options that impact behavior of Compactions:
<ul>
<p>
<li><code>Options::compaction_style</code> - RocksDB currently supports two
compaction algorithms - Universal  style and Level style. This option switches
between the two.  Can be kCompactionStyleUniversal or kCompactionStyleLevel.
If this is kCompactionStyleUniversal, then you can configure universal style
parameters with <code>Options::compaction_options_universal</code>.
<p>
<li><code>Options::disable_auto_compactions</code> - Disable automatic compactions.
Manual compactions can still be issued on this database.
<p>
<li><code>Options::compaction_filter</code> - Allows an application to modify/delete
a key-value during background compaction. The client must provide
compaction_filter_factory if it requires a new compaction filter to be used
for different compaction processes. Client should specify only one of filter
or factory.
<p>
<li><code>Options::compaction_filter_factory</code> - a factory that provides
compaction filter objects which allow an application to modify/delete a
key-value during background compaction.
</ul>
<p>
Other options impacting performance of compactions and when they get triggered
are: 
<ul>
<p>
<li> <code>Options::access_hint_on_compaction_start</code> - Specify the file access 
pattern once a compaction is started. It will be applied to all input files of a compaction. Default: NORMAL
<p>
<li> <code>Options::level0_file_num_compaction_trigger</code> -  Number of files to trigger level-0 compaction. 
A negative value means that level-0 compaction will not be triggered by number of files at all.
<p>
<li> <code>Options::max_mem_compaction_level</code> -  Maximum level to which a new compacted memtable is pushed if it
does not create overlap.  We try to push to level 2 to avoid the relatively expensive level 0=>1 compactions and to avoid some
expensive manifest file operations.  We do not push all the way to the largest level since that can generate a lot of wasted disk
space if the same key space is being repeatedly overwritten.
<p>
<li> <code>Options::target_file_size_base</code> and <code>Options::target_file_size_multiplier</code> - 
Target file size for compaction.  target_file_size_base is per-file size for level-1.
Target file size for level L can be calculated by target_file_size_base * (target_file_size_multiplier ^ (L-1))
For example, if target_file_size_base is 2MB and target_file_size_multiplier is 10, then each file on level-1 will
be 2MB, and each file on level 2 will be 20MB, and each file on level-3 will be 200MB. Default target_file_size_base is 2MB
and default target_file_size_multiplier is 1.
<p>
<li> <code>Options::expanded_compaction_factor</code> -  Maximum number of bytes in all compacted files.  We avoid expanding
the lower level file set of a compaction if it would make the total compaction cover more than
(expanded_compaction_factor * targetFileSizeLevel()) many bytes.
<p>
<li> <code>Options::source_compaction_factor</code> -    Maximum number of bytes in all source files to be compacted in a
single compaction run. We avoid picking too many files in the source level so that we do not exceed the total source bytes
for compaction to exceed (source_compaction_factor * targetFileSizeLevel()) many bytes.
Default:1, i.e. pick maxfilesize amount of data as the source of a compaction.
<p>
<li> <code>Options::max_grandparent_overlap_factor</code> -   Control maximum bytes of overlaps in grandparent (i.e., level+2) before we
stop building a single file in a level->level+1 compaction.
<p>
<li> <code>Options::disable_seek_compaction</code> -  Disable compaction triggered by seek.
With bloomfilter and fast storage, a miss on one level is very cheap if the file handle is cached in table cache
(which is true if max_open_files is large).
<p>
<li> <code>Options::max_background_compactions</code> - Maximum number of concurrent background jobs, submitted to
the default LOW priority thread pool
</ul>

<p>
You can learn more about all of those options in <code>rocksdb/options.h</code>

<h2> Universal style compaction specific settings</h2>
<p>
If you're using Universal style compaction, there is an object <code>CompactionOptionsUniversal</code>
that hold all the different options for that compaction. The exact definition is in
<code>rocksdb/universal_compaction.h</code> and you can set it in <code>Options::compaction_options_universal</code>.
Here we give short overview of options in <code>CompactionOptionsUniversal</code>:
<ul>
<p>
<li> <code>CompactionOptionsUniversal::size_ratio</code> - Percentage flexibilty while comparing file size. If the candidate file(s)
   size is 1% smaller than the next file's size, then include next file into
   this candidate set.  Default: 1
<p>
<li> <code>CompactionOptionsUniversal::min_merge_width</code> - The minimum number of files in a single compaction run. Default: 2
<p>
<li> <code>CompactionOptionsUniversal::max_merge_width</code> - The maximum number of files in a single compaction run. Default: UINT_MAX
<p>
<li> <code>CompactionOptionsUniversal::max_size_amplification_percent</code> - The size amplification is defined as the amount (in percentage) of
additional storage needed to store a single byte of data in the database.  For example, a size amplification of 2% means that a database that
contains 100 bytes of user-data may occupy upto 102 bytes of physical storage. By this definition, a fully compacted database has
a size amplification of 0%. Rocksdb uses the following heuristic to calculate size amplification: it assumes that all files excluding
the earliest file contribute to the size amplification.  Default: 200, which means that a 100 byte database could require upto
300 bytes of storage.
<p>
<li> <code>CompactionOptionsUniversal::compression_size_percent</code> - If this option is set to be -1 (the default value), all the output files
will follow compression type specified.  If this option is not negative, we will try to make sure compressed
size is just above this value. In normal cases, at least this percentage
of data will be compressed.
When we are compacting to a new file, here is the criteria whether
it needs to be compressed: assuming here are the list of files sorted
by generation time: [ A1...An B1...Bm C1...Ct ],
where A1 is the newest and Ct is the oldest, and we are going to compact
B1...Bm, we calculate the total size of all the files as total_size, as
well as  the total size of C1...Ct as total_C, the compaction output file
will be compressed iff total_C / total_size < this percentage
<p>
<li> <code>CompactionOptionsUniversal::stop_style</code> - The algorithm used to stop picking files into a single compaction run.
Can be kCompactionStopStyleSimilarSize (pick files of similar size) or kCompactionStopStyleTotalSize (total size of picked files > next file).
Default: kCompactionStopStyleTotalSize
</ul>

<h1>Thread pools</h1>
<p>
A thread pool is associated with Env environment object. The client has to create a thread pool by setting the number of background
threads using method <code>Env::SetBackgroundThreads()</code> defined in <code>rocksdb/env.h</code>.
We use the thread pool for compactions and memtable flushes.
Since memtable flushes are in critical code path (stalling memtable flush can stall writes, increasing p99), we suggest 
having two thread pools - with priorities HIGH and LOW. Memtable flushes can be set up to be scheduled on HIGH thread pool.
There are two options available for configuration of background compactions and flushes:
<ul>
<p>
<li> <code>Options::max_background_compactions</code> - Maximum number of concurrent background jobs,
submitted to the default LOW priority thread pool
<p>
<li> <code>Options::max_background_flushes</code> - Maximum number of concurrent background memtable flush jobs, submitted to
the HIGH priority thread pool.  By default, all background jobs (major compaction and memtable flush) go
to the LOW priority pool. If this option is set to a positive number, memtable flush jobs will be submitted to the HIGH priority pool.
It is important when the same Env is shared by multiple db instances.  Without a separate pool, long running major compaction jobs could
potentially block memtable flush jobs of other db instances, leading to unnecessary Put stalls.
</ul>
<p>
<pre>
  #include "rocksdb/env.h"
  #include "rocksdb/db.h"

  auto env = rocksdb::Env::Default();
  env->SetBackgroundThreads(2, rocksdb::Env::LOW);
  env->SetBackgroundThreads(1, rocksdb::Env::HIGH);
  rocksdb::DB* db;
  rocksdb::Options options;
  options.env = env;
  options.max_background_compactions = 2;
  options.max_background_flushes = 1;
  rocksdb::Status status = rocksdb::DB::Open(options, "/tmp/testdb", &amp;db);
  assert(status.ok());
  ...
</pre>
<h1>Approximate Sizes</h1>
<p>
The <code>GetApproximateSizes</code> method can used to get the approximate
number of bytes of file system space used by one or more key ranges.
<p>
<pre>
   rocksdb::Range ranges[2];
   ranges[0] = rocksdb::Range("a", "c");
   ranges[1] = rocksdb::Range("x", "z");
   uint64_t sizes[2];
   rocksdb::Status s = db-&gt;GetApproximateSizes(ranges, 2, sizes);
</pre>
The preceding call will set <code>sizes[0]</code> to the approximate number of
bytes of file system space used by the key range <code>[a..c)</code> and
<code>sizes[1]</code> to the approximate number of bytes used by the key range
<code>[x..z)</code>.
<p>
<h1>Environment</h1>
<p>
All file operations (and other operating system calls) issued by the
<code>rocksdb</code> implementation are routed through a <code>rocksdb::Env</code> object.
Sophisticated clients may wish to provide their own <code>Env</code>
implementation to get better control.  For example, an application may
introduce artificial delays in the file IO paths to limit the impact
of <code>rocksdb</code> on other activities in the system.
<p>
<pre>
  class SlowEnv : public rocksdb::Env {
    .. implementation of the Env interface ...
  };

  SlowEnv env;
  rocksdb::Options options;
  options.env = &amp;env;
  Status s = rocksdb::DB::Open(options, ...);
</pre>
<h1>Porting</h1>
<p>
<code>rocksdb</code> may be ported to a new platform by providing platform
specific implementations of the types/methods/functions exported by
<code>rocksdb/port/port.h</code>.  See <code>rocksdb/port/port_example.h</code> for more
details.
<p>
In addition, the new platform may need a new default <code>rocksdb::Env</code>
implementation.  See <code>rocksdb/util/env_posix.h</code> for an example.

<h1>Statistics</h1>
<p>
To be able to efficiently tune your application, it is always helpful if you
have access to usage statistics. You can collect those statistics by setting
<code>Options::table_stats_collectors</code> or
<code>Options::statistics</code>. For more information, refer to
<code>rocksdb/table_stats.h</code> and <code>rocksdb/statistics.h</code>.
These should not add significant overhead to your application and we
recommend exporting them to other monitoring tools.

<h1>Purging WAL files</h1>
<p>
By default, old write-ahead logs are deleted automatically when they fall out
of scope and application doesn't need them anymore. There are options that
enable the user to archive the logs and then delete them lazily, either in
TTL fashion or based on size limit.

The options are <code>Options::WAL_ttl_seconds</code> and
<code>Options::WAL_size_limit_MB</code>. Here is how they can be used:
<ul>
<li>
<p>
If both set to 0, logs will be deleted asap and will never get into the archive.
<li>
<p>
If <code>WAL_ttl_seconds</code> 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
<code>WAL_size_limit_MB</code>, they will be deleted starting with the
earliest until size_limit is met. All empty files will be deleted.
<li>
<p>
If <code>WAL_ttl_seconds</code> is not 0 and WAL_size_limit_MB is 0, then
WAL files will be checked every <code>WAL_ttl_seconds / 2</code> and those
that are older than WAL_ttl_seconds will be deleted.
<li>
<p>
If both are not 0, WAL files will be checked every 10 min and both
checks will be performed with ttl being first.
</ul>

<h1>Other Information</h1>
<p>
Details about the <code>rocksdb</code> implementation may be found in
the following documents:
<ul>
<li> <a href="https://github.com/facebook/rocksdb/wiki/Rocksdb-Architecture-Guide">
  RocksDB Architecture Guide</a>
<li> <a href="https://github.com/facebook/rocksdb/wiki/Rocksdb-Table-Format">
    Format of an immutable Table file</a>
<li> <a href="log_format.txt">Format of a log file</a>
</ul>

</body>
</html>