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
transaction_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.
#ifndef ROCKSDB_LITE
#include "rocksdb/db.h"
#include "rocksdb/options.h"
#include "rocksdb/slice.h"
#include "rocksdb/utilities/transaction.h"
#include "rocksdb/utilities/transaction_db.h"
using namespace rocksdb;
std::string kDBPath = "/tmp/rocksdb_transaction_example";
int main() {
// open DB
Options options;
TransactionDBOptions txn_db_options;
options.create_if_missing = true;
TransactionDB* txn_db;
Status s = TransactionDB::Open(options, txn_db_options, kDBPath, &txn_db);
assert(s.ok());
WriteOptions write_options;
ReadOptions read_options;
TransactionOptions txn_options;
std::string value;
////////////////////////////////////////////////////////
//
// Simple Transaction Example ("Read Committed")
//
////////////////////////////////////////////////////////
// Start a transaction
Transaction* txn = txn_db->BeginTransaction(write_options);
assert(txn);
// Read a key in this transaction
s = txn->Get(read_options, "abc", &value);
assert(s.IsNotFound());
// Write a key in this transaction
s = txn->Put("abc", "def");
assert(s.ok());
// Read a key OUTSIDE this transaction. Does not affect txn.
s = txn_db->Get(read_options, "abc", &value);
// Write a key OUTSIDE of this transaction.
// Does not affect txn since this is an unrelated key. If we wrote key 'abc'
// here, the transaction would fail to commit.
s = txn_db->Put(write_options, "xyz", "zzz");
// Commit transaction
s = txn->Commit();
assert(s.ok());
delete txn;
////////////////////////////////////////////////////////
//
// "Repeatable Read" (Snapshot Isolation) Example
// -- Using a single Snapshot
//
////////////////////////////////////////////////////////
// Set a snapshot at start of transaction by setting set_snapshot=true
txn_options.set_snapshot = true;
txn = txn_db->BeginTransaction(write_options, txn_options);
const Snapshot* snapshot = txn->GetSnapshot();
// Write a key OUTSIDE of transaction
s = txn_db->Put(write_options, "abc", "xyz");
assert(s.ok());
// Attempt to read a key using the snapshot. This will fail since
// the previous write outside this txn conflicts with this read.
read_options.snapshot = snapshot;
s = txn->GetForUpdate(read_options, "abc", &value);
assert(s.IsBusy());
txn->Rollback();
delete txn;
// Clear snapshot from read options since it is no longer valid
read_options.snapshot = nullptr;
snapshot = nullptr;
////////////////////////////////////////////////////////
//
// "Read Committed" (Monotonic Atomic Views) Example
// --Using multiple Snapshots
//
////////////////////////////////////////////////////////
// In this example, we set the snapshot multiple times. This is probably
// only necessary if you have very strict isolation requirements to
// implement.
// Set a snapshot at start of transaction
txn_options.set_snapshot = true;
txn = txn_db->BeginTransaction(write_options, txn_options);
// Do some reads and writes to key "x"
read_options.snapshot = txn_db->GetSnapshot();
s = txn->Get(read_options, "x", &value);
txn->Put("x", "x");
// Do a write outside of the transaction to key "y"
s = txn_db->Put(write_options, "y", "y");
// Set a new snapshot in the transaction
txn->SetSnapshot();
txn->SetSavePoint();
read_options.snapshot = txn_db->GetSnapshot();
// Do some reads and writes to key "y"
// Since the snapshot was advanced, the write done outside of the
// transaction does not conflict.
s = txn->GetForUpdate(read_options, "y", &value);
txn->Put("y", "y");
// Decide we want to revert the last write from this transaction.
txn->RollbackToSavePoint();
// Commit.
s = txn->Commit();
assert(s.ok());
delete txn;
// Clear snapshot from read options since it is no longer valid
read_options.snapshot = nullptr;
// Cleanup
delete txn_db;
DestroyDB(kDBPath, options);
return 0;
}
#endif // ROCKSDB_LITE
![swh spinner](/static/img/swh-spinner.gif)
Computing file changes ...