Revision 499ebb3ab5ea4207950fc95acf102b8f58add1c5 authored by Maysam Yabandeh on 24 June 2017, 21:06:43 UTC, committed by Facebook Github Bot on 24 June 2017, 21:11:29 UTC
Summary:
Throughput: 46k tps in our sysbench settings (filling the details later)

The idea is to have the simplest change that gives us a reasonable boost
in 2PC throughput.

Major design changes:
1. The WAL file internal buffer is not flushed after each write. Instead
it is flushed before critical operations (WAL copy via fs) or when
FlushWAL is called by MySQL. Flushing the WAL buffer is also protected
via mutex_.
2. Use two sequence numbers: last seq, and last seq for write. Last seq
is the last visible sequence number for reads. Last seq for write is the
next sequence number that should be used to write to WAL/memtable. This
allows to have a memtable write be in parallel to WAL writes.
3. BatchGroup is not used for writes. This means that we can have
parallel writers which changes a major assumption in the code base. To
accommodate for that i) allow only 1 WriteImpl that intends to write to
memtable via mem_mutex_--which is fine since in 2PC almost all of the memtable writes
come via group commit phase which is serial anyway, ii) make all the
parts in the code base that assumed to be the only writer (via
EnterUnbatched) to also acquire mem_mutex_, iii) stat updates are
protected via a stat_mutex_.

Note: the first commit has the approach figured out but is not clean.
Submitting the PR anyway to get the early feedback on the approach. If
we are ok with the approach I will go ahead with this updates:
0) Rebase with Yi's pipelining changes
1) Currently batching is disabled by default to make sure that it will be
consistent with all unit tests. Will make this optional via a config.
2) A couple of unit tests are disabled. They need to be updated with the
serial commit of 2PC taken into account.
3) Replacing BatchGroup with mem_mutex_ got a bit ugly as it requires
releasing mutex_ beforehand (the same way EnterUnbatched does). This
needs to be cleaned up.
Closes https://github.com/facebook/rocksdb/pull/2345

Differential Revision: D5210732

Pulled By: maysamyabandeh

fbshipit-source-id: 78653bd95a35cd1e831e555e0e57bdfd695355a4
1 parent 0ac4afb
Raw File
options_file_example.cc
// Copyright (c) 2011-present, Facebook, Inc.  All rights reserved.
// This source code is licensed under the BSD-style license found in the
// LICENSE file in the root directory of this source tree. An additional grant
// of patent rights can be found in the PATENTS file in the same directory.
//
// This file demonstrates how to use the utility functions defined in
// rocksdb/utilities/options_util.h to open a rocksdb database without
// remembering all the rocksdb options.
#include <cstdio>
#include <string>
#include <vector>

#include "rocksdb/cache.h"
#include "rocksdb/compaction_filter.h"
#include "rocksdb/db.h"
#include "rocksdb/options.h"
#include "rocksdb/slice.h"
#include "rocksdb/table.h"
#include "rocksdb/utilities/options_util.h"

using namespace rocksdb;

std::string kDBPath = "/tmp/rocksdb_options_file_example";

namespace {
// A dummy compaction filter
class DummyCompactionFilter : public CompactionFilter {
 public:
  virtual ~DummyCompactionFilter() {}
  virtual bool Filter(int level, const Slice& key, const Slice& existing_value,
                      std::string* new_value, bool* value_changed) const {
    return false;
  }
  virtual const char* Name() const { return "DummyCompactionFilter"; }
};

}  // namespace

int main() {
  DBOptions db_opt;
  db_opt.create_if_missing = true;

  std::vector<ColumnFamilyDescriptor> cf_descs;
  cf_descs.push_back({kDefaultColumnFamilyName, ColumnFamilyOptions()});
  cf_descs.push_back({"new_cf", ColumnFamilyOptions()});

  // initialize BlockBasedTableOptions
  auto cache = NewLRUCache(1 * 1024 * 1024 * 1024);
  BlockBasedTableOptions bbt_opts;
  bbt_opts.block_size = 32 * 1024;
  bbt_opts.block_cache = cache;

  // initialize column families options
  std::unique_ptr<CompactionFilter> compaction_filter;
  compaction_filter.reset(new DummyCompactionFilter());
  cf_descs[0].options.table_factory.reset(NewBlockBasedTableFactory(bbt_opts));
  cf_descs[0].options.compaction_filter = compaction_filter.get();
  cf_descs[1].options.table_factory.reset(NewBlockBasedTableFactory(bbt_opts));

  // destroy and open DB
  DB* db;
  Status s = DestroyDB(kDBPath, Options(db_opt, cf_descs[0].options));
  assert(s.ok());
  s = DB::Open(Options(db_opt, cf_descs[0].options), kDBPath, &db);
  assert(s.ok());

  // Create column family, and rocksdb will persist the options.
  ColumnFamilyHandle* cf;
  s = db->CreateColumnFamily(ColumnFamilyOptions(), "new_cf", &cf);
  assert(s.ok());

  // close DB
  delete cf;
  delete db;

  // In the following code, we will reopen the rocksdb instance using
  // the options file stored in the db directory.

  // Load the options file.
  DBOptions loaded_db_opt;
  std::vector<ColumnFamilyDescriptor> loaded_cf_descs;
  s = LoadLatestOptions(kDBPath, Env::Default(), &loaded_db_opt,
                        &loaded_cf_descs);
  assert(s.ok());
  assert(loaded_db_opt.create_if_missing == db_opt.create_if_missing);

  // Initialize pointer options for each column family
  for (size_t i = 0; i < loaded_cf_descs.size(); ++i) {
    auto* loaded_bbt_opt = reinterpret_cast<BlockBasedTableOptions*>(
        loaded_cf_descs[0].options.table_factory->GetOptions());
    // Expect the same as BlockBasedTableOptions will be loaded form file.
    assert(loaded_bbt_opt->block_size == bbt_opts.block_size);
    // However, block_cache needs to be manually initialized as documented
    // in rocksdb/utilities/options_util.h.
    loaded_bbt_opt->block_cache = cache;
  }
  // In addition, as pointer options are initialized with default value,
  // we need to properly initialized all the pointer options if non-defalut
  // values are used before calling DB::Open().
  assert(loaded_cf_descs[0].options.compaction_filter == nullptr);
  loaded_cf_descs[0].options.compaction_filter = compaction_filter.get();

  // reopen the db using the loaded options.
  std::vector<ColumnFamilyHandle*> handles;
  s = DB::Open(loaded_db_opt, kDBPath, loaded_cf_descs, &handles, &db);
  assert(s.ok());

  // close DB
  for (auto* handle : handles) {
    delete handle;
  }
  delete db;
}
back to top