2016-02-10 00:12:00 +01:00
|
|
|
// Copyright (c) 2011-present, Facebook, Inc. All rights reserved.
|
2017-07-16 01:03:42 +02:00
|
|
|
// This source code is licensed under both the GPLv2 (found in the
|
|
|
|
// COPYING file in the root directory) and Apache 2.0 License
|
|
|
|
// (found in the LICENSE.Apache file in the root directory).
|
2017-04-06 04:02:00 +02:00
|
|
|
#include "options/options_helper.h"
|
2014-09-17 21:46:32 +02:00
|
|
|
|
|
|
|
#include <cassert>
|
2014-10-10 19:00:12 +02:00
|
|
|
#include <cctype>
|
2015-04-24 04:17:57 +02:00
|
|
|
#include <cstdlib>
|
2014-10-02 01:19:16 +02:00
|
|
|
#include <unordered_set>
|
2015-10-03 00:35:32 +02:00
|
|
|
#include <vector>
|
2019-05-30 23:47:29 +02:00
|
|
|
|
2014-12-22 22:18:57 +01:00
|
|
|
#include "rocksdb/cache.h"
|
2015-10-03 00:35:32 +02:00
|
|
|
#include "rocksdb/compaction_filter.h"
|
2015-07-15 23:51:51 +02:00
|
|
|
#include "rocksdb/convenience.h"
|
2014-12-22 22:18:57 +01:00
|
|
|
#include "rocksdb/filter_policy.h"
|
2015-10-03 00:35:32 +02:00
|
|
|
#include "rocksdb/memtablerep.h"
|
|
|
|
#include "rocksdb/merge_operator.h"
|
2014-09-17 21:46:32 +02:00
|
|
|
#include "rocksdb/options.h"
|
2015-03-06 23:21:15 +01:00
|
|
|
#include "rocksdb/rate_limiter.h"
|
2015-01-16 00:33:12 +01:00
|
|
|
#include "rocksdb/slice_transform.h"
|
2014-12-22 22:18:57 +01:00
|
|
|
#include "rocksdb/table.h"
|
2019-03-26 22:15:26 +01:00
|
|
|
#include "rocksdb/utilities/object_registry.h"
|
2019-05-30 23:47:29 +02:00
|
|
|
#include "table/block_based/block_based_table_factory.h"
|
2019-05-31 02:39:43 +02:00
|
|
|
#include "table/plain/plain_table_factory.h"
|
2017-07-29 01:23:50 +02:00
|
|
|
#include "util/cast_util.h"
|
2015-08-18 22:30:18 +02:00
|
|
|
#include "util/string_util.h"
|
2014-09-17 21:46:32 +02:00
|
|
|
|
|
|
|
namespace rocksdb {
|
|
|
|
|
2016-09-24 01:34:04 +02:00
|
|
|
DBOptions BuildDBOptions(const ImmutableDBOptions& immutable_db_options,
|
|
|
|
const MutableDBOptions& mutable_db_options) {
|
|
|
|
DBOptions options;
|
|
|
|
|
|
|
|
options.create_if_missing = immutable_db_options.create_if_missing;
|
|
|
|
options.create_missing_column_families =
|
|
|
|
immutable_db_options.create_missing_column_families;
|
|
|
|
options.error_if_exists = immutable_db_options.error_if_exists;
|
|
|
|
options.paranoid_checks = immutable_db_options.paranoid_checks;
|
|
|
|
options.env = immutable_db_options.env;
|
|
|
|
options.rate_limiter = immutable_db_options.rate_limiter;
|
|
|
|
options.sst_file_manager = immutable_db_options.sst_file_manager;
|
|
|
|
options.info_log = immutable_db_options.info_log;
|
|
|
|
options.info_log_level = immutable_db_options.info_log_level;
|
2017-05-04 05:46:17 +02:00
|
|
|
options.max_open_files = mutable_db_options.max_open_files;
|
2016-09-24 01:34:04 +02:00
|
|
|
options.max_file_opening_threads =
|
|
|
|
immutable_db_options.max_file_opening_threads;
|
2016-11-15 07:45:16 +01:00
|
|
|
options.max_total_wal_size = mutable_db_options.max_total_wal_size;
|
2016-09-24 01:34:04 +02:00
|
|
|
options.statistics = immutable_db_options.statistics;
|
|
|
|
options.use_fsync = immutable_db_options.use_fsync;
|
|
|
|
options.db_paths = immutable_db_options.db_paths;
|
|
|
|
options.db_log_dir = immutable_db_options.db_log_dir;
|
|
|
|
options.wal_dir = immutable_db_options.wal_dir;
|
|
|
|
options.delete_obsolete_files_period_micros =
|
2016-12-05 23:09:35 +01:00
|
|
|
mutable_db_options.delete_obsolete_files_period_micros;
|
2017-05-24 20:25:38 +02:00
|
|
|
options.max_background_jobs = mutable_db_options.max_background_jobs;
|
2016-09-24 01:34:04 +02:00
|
|
|
options.base_background_compactions =
|
2016-10-14 21:25:39 +02:00
|
|
|
mutable_db_options.base_background_compactions;
|
2016-09-24 01:34:04 +02:00
|
|
|
options.max_background_compactions =
|
2016-10-14 21:25:39 +02:00
|
|
|
mutable_db_options.max_background_compactions;
|
2017-09-28 02:37:08 +02:00
|
|
|
options.bytes_per_sync = mutable_db_options.bytes_per_sync;
|
|
|
|
options.wal_bytes_per_sync = mutable_db_options.wal_bytes_per_sync;
|
Optionally wait on bytes_per_sync to smooth I/O (#5183)
Summary:
The existing implementation does not guarantee bytes reach disk every `bytes_per_sync` when writing SST files, or every `wal_bytes_per_sync` when writing WALs. This can cause confusing behavior for users who enable this feature to avoid large syncs during flush and compaction, but then end up hitting them anyways.
My understanding of the existing behavior is we used `sync_file_range` with `SYNC_FILE_RANGE_WRITE` to submit ranges for async writeback, such that we could continue processing the next range of bytes while that I/O is happening. I believe we can preserve that benefit while also limiting how far the processing can get ahead of the I/O, which prevents huge syncs from happening when the file finishes.
Consider this `sync_file_range` usage: `sync_file_range(fd_, 0, static_cast<off_t>(offset + nbytes), SYNC_FILE_RANGE_WAIT_BEFORE | SYNC_FILE_RANGE_WRITE)`. Expanding the range to start at 0 and adding the `SYNC_FILE_RANGE_WAIT_BEFORE` flag causes any pending writeback (like from a previous call to `sync_file_range`) to finish before it proceeds to submit the latest `nbytes` for writeback. The latest `nbytes` are still written back asynchronously, unless processing exceeds I/O speed, in which case the following `sync_file_range` will need to wait on it.
There is a second change in this PR to use `fdatasync` when `sync_file_range` is unavailable (determined statically) or has some known problem with the underlying filesystem (determined dynamically).
The above two changes only apply when the user enables a new option, `strict_bytes_per_sync`.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5183
Differential Revision: D14953553
Pulled By: siying
fbshipit-source-id: 445c3862e019fb7b470f9c7f314fc231b62706e9
2019-04-22 20:48:45 +02:00
|
|
|
options.strict_bytes_per_sync = mutable_db_options.strict_bytes_per_sync;
|
2016-09-24 01:34:04 +02:00
|
|
|
options.max_subcompactions = immutable_db_options.max_subcompactions;
|
|
|
|
options.max_background_flushes = immutable_db_options.max_background_flushes;
|
|
|
|
options.max_log_file_size = immutable_db_options.max_log_file_size;
|
|
|
|
options.log_file_time_to_roll = immutable_db_options.log_file_time_to_roll;
|
|
|
|
options.keep_log_file_num = immutable_db_options.keep_log_file_num;
|
|
|
|
options.recycle_log_file_num = immutable_db_options.recycle_log_file_num;
|
|
|
|
options.max_manifest_file_size = immutable_db_options.max_manifest_file_size;
|
|
|
|
options.table_cache_numshardbits =
|
|
|
|
immutable_db_options.table_cache_numshardbits;
|
|
|
|
options.WAL_ttl_seconds = immutable_db_options.wal_ttl_seconds;
|
|
|
|
options.WAL_size_limit_MB = immutable_db_options.wal_size_limit_mb;
|
|
|
|
options.manifest_preallocation_size =
|
|
|
|
immutable_db_options.manifest_preallocation_size;
|
|
|
|
options.allow_mmap_reads = immutable_db_options.allow_mmap_reads;
|
|
|
|
options.allow_mmap_writes = immutable_db_options.allow_mmap_writes;
|
2016-10-28 19:36:05 +02:00
|
|
|
options.use_direct_reads = immutable_db_options.use_direct_reads;
|
2017-04-13 22:07:33 +02:00
|
|
|
options.use_direct_io_for_flush_and_compaction =
|
|
|
|
immutable_db_options.use_direct_io_for_flush_and_compaction;
|
2016-09-24 01:34:04 +02:00
|
|
|
options.allow_fallocate = immutable_db_options.allow_fallocate;
|
|
|
|
options.is_fd_close_on_exec = immutable_db_options.is_fd_close_on_exec;
|
2017-03-21 06:50:56 +01:00
|
|
|
options.stats_dump_period_sec = mutable_db_options.stats_dump_period_sec;
|
2019-02-21 00:46:59 +01:00
|
|
|
options.stats_persist_period_sec =
|
|
|
|
mutable_db_options.stats_persist_period_sec;
|
2019-06-18 00:17:43 +02:00
|
|
|
options.persist_stats_to_disk = immutable_db_options.persist_stats_to_disk;
|
2019-02-21 00:46:59 +01:00
|
|
|
options.stats_history_buffer_size =
|
|
|
|
mutable_db_options.stats_history_buffer_size;
|
2016-09-24 01:34:04 +02:00
|
|
|
options.advise_random_on_open = immutable_db_options.advise_random_on_open;
|
|
|
|
options.db_write_buffer_size = immutable_db_options.db_write_buffer_size;
|
|
|
|
options.write_buffer_manager = immutable_db_options.write_buffer_manager;
|
|
|
|
options.access_hint_on_compaction_start =
|
|
|
|
immutable_db_options.access_hint_on_compaction_start;
|
|
|
|
options.new_table_reader_for_compaction_inputs =
|
|
|
|
immutable_db_options.new_table_reader_for_compaction_inputs;
|
|
|
|
options.compaction_readahead_size =
|
2017-11-17 02:46:43 +01:00
|
|
|
mutable_db_options.compaction_readahead_size;
|
2016-09-24 01:34:04 +02:00
|
|
|
options.random_access_max_buffer_size =
|
|
|
|
immutable_db_options.random_access_max_buffer_size;
|
|
|
|
options.writable_file_max_buffer_size =
|
2017-10-31 21:49:25 +01:00
|
|
|
mutable_db_options.writable_file_max_buffer_size;
|
2016-09-24 01:34:04 +02:00
|
|
|
options.use_adaptive_mutex = immutable_db_options.use_adaptive_mutex;
|
|
|
|
options.listeners = immutable_db_options.listeners;
|
|
|
|
options.enable_thread_tracking = immutable_db_options.enable_thread_tracking;
|
2016-11-13 00:43:33 +01:00
|
|
|
options.delayed_write_rate = mutable_db_options.delayed_write_rate;
|
2018-03-13 19:50:16 +01:00
|
|
|
options.enable_pipelined_write = immutable_db_options.enable_pipelined_write;
|
2019-05-14 02:43:47 +02:00
|
|
|
options.unordered_write = immutable_db_options.unordered_write;
|
2016-09-24 01:34:04 +02:00
|
|
|
options.allow_concurrent_memtable_write =
|
|
|
|
immutable_db_options.allow_concurrent_memtable_write;
|
|
|
|
options.enable_write_thread_adaptive_yield =
|
|
|
|
immutable_db_options.enable_write_thread_adaptive_yield;
|
|
|
|
options.write_thread_max_yield_usec =
|
|
|
|
immutable_db_options.write_thread_max_yield_usec;
|
|
|
|
options.write_thread_slow_yield_usec =
|
|
|
|
immutable_db_options.write_thread_slow_yield_usec;
|
|
|
|
options.skip_stats_update_on_db_open =
|
|
|
|
immutable_db_options.skip_stats_update_on_db_open;
|
|
|
|
options.wal_recovery_mode = immutable_db_options.wal_recovery_mode;
|
|
|
|
options.allow_2pc = immutable_db_options.allow_2pc;
|
|
|
|
options.row_cache = immutable_db_options.row_cache;
|
|
|
|
#ifndef ROCKSDB_LITE
|
|
|
|
options.wal_filter = immutable_db_options.wal_filter;
|
|
|
|
#endif // ROCKSDB_LITE
|
|
|
|
options.fail_if_options_file_error =
|
|
|
|
immutable_db_options.fail_if_options_file_error;
|
|
|
|
options.dump_malloc_stats = immutable_db_options.dump_malloc_stats;
|
|
|
|
options.avoid_flush_during_recovery =
|
|
|
|
immutable_db_options.avoid_flush_during_recovery;
|
2016-11-02 23:22:13 +01:00
|
|
|
options.avoid_flush_during_shutdown =
|
|
|
|
mutable_db_options.avoid_flush_during_shutdown;
|
2017-05-17 20:32:26 +02:00
|
|
|
options.allow_ingest_behind =
|
|
|
|
immutable_db_options.allow_ingest_behind;
|
Added support for differential snapshots
Summary:
The motivation for this PR is to add to RocksDB support for differential (incremental) snapshots, as snapshot of the DB changes between two points in time (one can think of it as diff between to sequence numbers, or the diff D which can be thought of as an SST file or just set of KVs that can be applied to sequence number S1 to get the database to the state at sequence number S2).
This feature would be useful for various distributed storages layers built on top of RocksDB, as it should help reduce resources (time and network bandwidth) needed to recover and rebuilt DB instances as replicas in the context of distributed storages.
From the API standpoint that would like client app requesting iterator between (start seqnum) and current DB state, and reading the "diff".
This is a very draft PR for initial review in the discussion on the approach, i'm going to rework some parts and keep updating the PR.
For now, what's done here according to initial discussions:
Preserving deletes:
- We want to be able to optionally preserve recent deletes for some defined period of time, so that if a delete came in recently and might need to be included in the next incremental snapshot it would't get dropped by a compaction. This is done by adding new param to Options (preserve deletes flag) and new variable to DB Impl where we keep track of the sequence number after which we don't want to drop tombstones, even if they are otherwise eligible for deletion.
- I also added a new API call for clients to be able to advance this cutoff seqnum after which we drop deletes; i assume it's more flexible to let clients control this, since otherwise we'd need to keep some kind of timestamp < -- > seqnum mapping inside the DB, which sounds messy and painful to support. Clients could make use of it by periodically calling GetLatestSequenceNumber(), noting the timestamp, doing some calculation and figuring out by how much we need to advance the cutoff seqnum.
- Compaction codepath in compaction_iterator.cc has been modified to avoid dropping tombstones with seqnum > cutoff seqnum.
Iterator changes:
- couple params added to ReadOptions, to optionally allow client to request internal keys instead of user keys (so that client can get the latest value of a key, be it delete marker or a put), as well as min timestamp and min seqnum.
TableCache changes:
- I modified table_cache code to be able to quickly exclude SST files from iterators heep if creation_time on the file is less then iter_start_ts as passed in ReadOptions. That would help a lot in some DB settings (like reading very recent data only or using FIFO compactions), but not so much for universal compaction with more or less long iterator time span.
What's left:
- Still looking at how to best plug that inside DBIter codepath. So far it seems that FindNextUserKeyInternal only parses values as UserKeys, and iter->key() call generally returns user key. Can we add new API to DBIter as internal_key(), and modify this internal method to optionally set saved_key_ to point to the full internal key? I don't need to store actual seqnum there, but I do need to store type.
Closes https://github.com/facebook/rocksdb/pull/2999
Differential Revision: D6175602
Pulled By: mikhail-antonov
fbshipit-source-id: c779a6696ee2d574d86c69cec866a3ae095aa900
2017-11-02 02:43:29 +01:00
|
|
|
options.preserve_deletes =
|
|
|
|
immutable_db_options.preserve_deletes;
|
2018-05-14 19:53:32 +02:00
|
|
|
options.two_write_queues = immutable_db_options.two_write_queues;
|
|
|
|
options.manual_wal_flush = immutable_db_options.manual_wal_flush;
|
2018-11-12 21:22:10 +01:00
|
|
|
options.atomic_flush = immutable_db_options.atomic_flush;
|
2019-04-02 02:07:38 +02:00
|
|
|
options.avoid_unnecessary_blocking_io =
|
|
|
|
immutable_db_options.avoid_unnecessary_blocking_io;
|
2019-07-19 20:54:38 +02:00
|
|
|
options.log_readahead_size = immutable_db_options.log_readahead_size;
|
2016-09-24 01:34:04 +02:00
|
|
|
return options;
|
|
|
|
}
|
|
|
|
|
2016-09-15 07:10:28 +02:00
|
|
|
ColumnFamilyOptions BuildColumnFamilyOptions(
|
|
|
|
const ColumnFamilyOptions& options,
|
|
|
|
const MutableCFOptions& mutable_cf_options) {
|
|
|
|
ColumnFamilyOptions cf_opts(options);
|
|
|
|
|
|
|
|
// Memtable related options
|
|
|
|
cf_opts.write_buffer_size = mutable_cf_options.write_buffer_size;
|
|
|
|
cf_opts.max_write_buffer_number = mutable_cf_options.max_write_buffer_number;
|
|
|
|
cf_opts.arena_block_size = mutable_cf_options.arena_block_size;
|
|
|
|
cf_opts.memtable_prefix_bloom_size_ratio =
|
|
|
|
mutable_cf_options.memtable_prefix_bloom_size_ratio;
|
2019-02-19 21:12:25 +01:00
|
|
|
cf_opts.memtable_whole_key_filtering =
|
|
|
|
mutable_cf_options.memtable_whole_key_filtering;
|
2016-09-15 07:10:28 +02:00
|
|
|
cf_opts.memtable_huge_page_size = mutable_cf_options.memtable_huge_page_size;
|
|
|
|
cf_opts.max_successive_merges = mutable_cf_options.max_successive_merges;
|
|
|
|
cf_opts.inplace_update_num_locks =
|
|
|
|
mutable_cf_options.inplace_update_num_locks;
|
2018-05-21 23:33:55 +02:00
|
|
|
cf_opts.prefix_extractor = mutable_cf_options.prefix_extractor;
|
2016-09-15 07:10:28 +02:00
|
|
|
|
|
|
|
// Compaction related options
|
|
|
|
cf_opts.disable_auto_compactions =
|
|
|
|
mutable_cf_options.disable_auto_compactions;
|
2018-05-08 21:06:01 +02:00
|
|
|
cf_opts.soft_pending_compaction_bytes_limit =
|
|
|
|
mutable_cf_options.soft_pending_compaction_bytes_limit;
|
|
|
|
cf_opts.hard_pending_compaction_bytes_limit =
|
|
|
|
mutable_cf_options.hard_pending_compaction_bytes_limit;
|
2016-09-15 07:10:28 +02:00
|
|
|
cf_opts.level0_file_num_compaction_trigger =
|
|
|
|
mutable_cf_options.level0_file_num_compaction_trigger;
|
|
|
|
cf_opts.level0_slowdown_writes_trigger =
|
|
|
|
mutable_cf_options.level0_slowdown_writes_trigger;
|
|
|
|
cf_opts.level0_stop_writes_trigger =
|
|
|
|
mutable_cf_options.level0_stop_writes_trigger;
|
|
|
|
cf_opts.max_compaction_bytes = mutable_cf_options.max_compaction_bytes;
|
|
|
|
cf_opts.target_file_size_base = mutable_cf_options.target_file_size_base;
|
|
|
|
cf_opts.target_file_size_multiplier =
|
|
|
|
mutable_cf_options.target_file_size_multiplier;
|
|
|
|
cf_opts.max_bytes_for_level_base =
|
|
|
|
mutable_cf_options.max_bytes_for_level_base;
|
2019-05-04 02:26:20 +02:00
|
|
|
cf_opts.snap_refresh_nanos = mutable_cf_options.snap_refresh_nanos;
|
2016-09-15 07:10:28 +02:00
|
|
|
cf_opts.max_bytes_for_level_multiplier =
|
|
|
|
mutable_cf_options.max_bytes_for_level_multiplier;
|
2018-07-16 23:24:33 +02:00
|
|
|
cf_opts.ttl = mutable_cf_options.ttl;
|
Periodic Compactions (#5166)
Summary:
Introducing Periodic Compactions.
This feature allows all the files in a CF to be periodically compacted. It could help in catching any corruptions that could creep into the DB proactively as every file is constantly getting re-compacted. And also, of course, it helps to cleanup data older than certain threshold.
- Introduced a new option `periodic_compaction_time` to control how long a file can live without being compacted in a CF.
- This works across all levels.
- The files are put in the same level after going through the compaction. (Related files in the same level are picked up as `ExpandInputstoCleanCut` is used).
- Compaction filters, if any, are invoked as usual.
- A new table property, `file_creation_time`, is introduced to implement this feature. This property is set to the time at which the SST file was created (and that time is given by the underlying Env/OS).
This feature can be enabled on its own, or in conjunction with `ttl`. It is possible to set a different time threshold for the bottom level when used in conjunction with ttl. Since `ttl` works only on 0 to last but one levels, you could set `ttl` to, say, 1 day, and `periodic_compaction_time` to, say, 7 days. Since `ttl < periodic_compaction_time` all files in last but one levels keep getting picked up based on ttl, and almost never based on periodic_compaction_time. The files in the bottom level get picked up for compaction based on `periodic_compaction_time`.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5166
Differential Revision: D14884441
Pulled By: sagar0
fbshipit-source-id: 408426cbacb409c06386a98632dcf90bfa1bda47
2019-04-11 04:24:25 +02:00
|
|
|
cf_opts.periodic_compaction_seconds =
|
|
|
|
mutable_cf_options.periodic_compaction_seconds;
|
2016-09-15 07:10:28 +02:00
|
|
|
|
|
|
|
cf_opts.max_bytes_for_level_multiplier_additional.clear();
|
|
|
|
for (auto value :
|
|
|
|
mutable_cf_options.max_bytes_for_level_multiplier_additional) {
|
|
|
|
cf_opts.max_bytes_for_level_multiplier_additional.emplace_back(value);
|
|
|
|
}
|
|
|
|
|
2017-10-20 00:19:20 +02:00
|
|
|
cf_opts.compaction_options_fifo = mutable_cf_options.compaction_options_fifo;
|
2017-12-11 22:12:12 +01:00
|
|
|
cf_opts.compaction_options_universal =
|
|
|
|
mutable_cf_options.compaction_options_universal;
|
2017-10-20 00:19:20 +02:00
|
|
|
|
2016-09-15 07:10:28 +02:00
|
|
|
// Misc options
|
|
|
|
cf_opts.max_sequential_skip_in_iterations =
|
|
|
|
mutable_cf_options.max_sequential_skip_in_iterations;
|
|
|
|
cf_opts.paranoid_file_checks = mutable_cf_options.paranoid_file_checks;
|
|
|
|
cf_opts.report_bg_io_stats = mutable_cf_options.report_bg_io_stats;
|
|
|
|
cf_opts.compression = mutable_cf_options.compression;
|
2019-03-18 20:07:35 +01:00
|
|
|
cf_opts.sample_for_compression = mutable_cf_options.sample_for_compression;
|
2016-09-15 07:10:28 +02:00
|
|
|
|
|
|
|
cf_opts.table_factory = options.table_factory;
|
|
|
|
// TODO(yhchiang): find some way to handle the following derived options
|
|
|
|
// * max_file_size
|
|
|
|
|
|
|
|
return cf_opts;
|
|
|
|
}
|
|
|
|
|
2017-11-18 02:02:13 +01:00
|
|
|
std::map<CompactionStyle, std::string>
|
|
|
|
OptionsHelper::compaction_style_to_string = {
|
|
|
|
{kCompactionStyleLevel, "kCompactionStyleLevel"},
|
|
|
|
{kCompactionStyleUniversal, "kCompactionStyleUniversal"},
|
|
|
|
{kCompactionStyleFIFO, "kCompactionStyleFIFO"},
|
|
|
|
{kCompactionStyleNone, "kCompactionStyleNone"}};
|
|
|
|
|
|
|
|
std::map<CompactionPri, std::string> OptionsHelper::compaction_pri_to_string = {
|
|
|
|
{kByCompensatedSize, "kByCompensatedSize"},
|
|
|
|
{kOldestLargestSeqFirst, "kOldestLargestSeqFirst"},
|
|
|
|
{kOldestSmallestSeqFirst, "kOldestSmallestSeqFirst"},
|
|
|
|
{kMinOverlappingRatio, "kMinOverlappingRatio"}};
|
|
|
|
|
|
|
|
std::map<CompactionStopStyle, std::string>
|
|
|
|
OptionsHelper::compaction_stop_style_to_string = {
|
|
|
|
{kCompactionStopStyleSimilarSize, "kCompactionStopStyleSimilarSize"},
|
|
|
|
{kCompactionStopStyleTotalSize, "kCompactionStopStyleTotalSize"}};
|
|
|
|
|
|
|
|
std::unordered_map<std::string, ChecksumType>
|
|
|
|
OptionsHelper::checksum_type_string_map = {{"kNoChecksum", kNoChecksum},
|
|
|
|
{"kCRC32c", kCRC32c},
|
2018-11-01 23:39:40 +01:00
|
|
|
{"kxxHash", kxxHash},
|
|
|
|
{"kxxHash64", kxxHash64}};
|
2017-11-18 02:02:13 +01:00
|
|
|
|
2018-01-23 23:36:54 +01:00
|
|
|
std::unordered_map<std::string, CompressionType>
|
|
|
|
OptionsHelper::compression_type_string_map = {
|
|
|
|
{"kNoCompression", kNoCompression},
|
|
|
|
{"kSnappyCompression", kSnappyCompression},
|
|
|
|
{"kZlibCompression", kZlibCompression},
|
|
|
|
{"kBZip2Compression", kBZip2Compression},
|
|
|
|
{"kLZ4Compression", kLZ4Compression},
|
|
|
|
{"kLZ4HCCompression", kLZ4HCCompression},
|
|
|
|
{"kXpressCompression", kXpressCompression},
|
|
|
|
{"kZSTD", kZSTD},
|
|
|
|
{"kZSTDNotFinalCompression", kZSTDNotFinalCompression},
|
|
|
|
{"kDisableCompressionOption", kDisableCompressionOption}};
|
2014-11-13 20:39:30 +01:00
|
|
|
#ifndef ROCKSDB_LITE
|
RocksDB Options file format and its serialization / deserialization.
Summary:
This patch defines the format of RocksDB options file, which
follows the INI file format, and implements functions for its
serialization and deserialization. An example RocksDB options
file can be found in examples/rocksdb_option_file_example.ini.
A typical RocksDB options file has three sections, which are
Version, DBOptions, and more than one CFOptions. The RocksDB
options file in general follows the basic INI file format
with the following extensions / modifications:
* Escaped characters
We escaped the following characters:
- \n -- line feed - new line
- \r -- carriage return
- \\ -- backslash \
- \: -- colon symbol :
- \# -- hash tag #
* Comments
We support # style comments. Comments can appear at the ending
part of a line.
* Statements
A statement is of the form option_name = value.
Each statement contains a '=', where extra white-spaces
are supported. However, we don't support multi-lined statement.
Furthermore, each line can only contain at most one statement.
* Section
Sections are of the form [SecitonTitle "SectionArgument"],
where section argument is optional.
* List
We use colon-separated string to represent a list.
For instance, n1:n2:n3:n4 is a list containing four values.
Below is an example of a RocksDB options file:
[Version]
rocksdb_version=4.0.0
options_file_version=1.0
[DBOptions]
max_open_files=12345
max_background_flushes=301
[CFOptions "default"]
[CFOptions "the second column family"]
[CFOptions "the third column family"]
Test Plan: Added many tests in options_test.cc
Reviewers: igor, IslamAbdelRahman, sdong, anthony
Reviewed By: anthony
Subscribers: maykov, dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D46059
2015-09-29 23:42:40 +02:00
|
|
|
|
2019-03-26 22:15:26 +01:00
|
|
|
const std::string kNameComparator = "comparator";
|
2019-04-25 20:31:58 +02:00
|
|
|
const std::string kNameEnv = "env";
|
2019-03-28 22:50:06 +01:00
|
|
|
const std::string kNameMergeOperator = "merge_operator";
|
2019-03-26 22:15:26 +01:00
|
|
|
|
2017-10-20 00:19:20 +02:00
|
|
|
template <typename T>
|
|
|
|
Status GetStringFromStruct(
|
|
|
|
std::string* opt_string, const T& options,
|
2019-05-15 23:19:04 +02:00
|
|
|
const std::unordered_map<std::string, OptionTypeInfo>& type_info,
|
2017-10-20 00:19:20 +02:00
|
|
|
const std::string& delimiter);
|
|
|
|
|
2016-08-17 09:42:35 +02:00
|
|
|
namespace {
|
2015-11-18 01:41:54 +01:00
|
|
|
template <typename T>
|
|
|
|
bool ParseEnum(const std::unordered_map<std::string, T>& type_map,
|
2015-10-30 23:58:46 +01:00
|
|
|
const std::string& type, T* value) {
|
|
|
|
auto iter = type_map.find(type);
|
|
|
|
if (iter != type_map.end()) {
|
|
|
|
*value = iter->second;
|
|
|
|
return true;
|
|
|
|
}
|
|
|
|
return false;
|
|
|
|
}
|
|
|
|
|
2015-11-18 01:41:54 +01:00
|
|
|
template <typename T>
|
|
|
|
bool SerializeEnum(const std::unordered_map<std::string, T>& type_map,
|
2015-10-30 23:58:46 +01:00
|
|
|
const T& type, std::string* value) {
|
|
|
|
for (const auto& pair : type_map) {
|
|
|
|
if (pair.second == type) {
|
|
|
|
*value = pair.first;
|
2015-10-03 00:35:32 +02:00
|
|
|
return true;
|
2015-10-30 23:58:46 +01:00
|
|
|
}
|
2015-10-03 00:35:32 +02:00
|
|
|
}
|
2015-10-30 23:58:46 +01:00
|
|
|
return false;
|
2015-10-03 00:35:32 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
bool SerializeVectorCompressionType(const std::vector<CompressionType>& types,
|
|
|
|
std::string* value) {
|
|
|
|
std::stringstream ss;
|
|
|
|
bool result;
|
|
|
|
for (size_t i = 0; i < types.size(); ++i) {
|
|
|
|
if (i > 0) {
|
|
|
|
ss << ':';
|
|
|
|
}
|
|
|
|
std::string string_type;
|
2015-10-30 23:58:46 +01:00
|
|
|
result = SerializeEnum<CompressionType>(compression_type_string_map,
|
2015-11-18 01:41:54 +01:00
|
|
|
types[i], &string_type);
|
2015-10-03 00:35:32 +02:00
|
|
|
if (result == false) {
|
|
|
|
return result;
|
|
|
|
}
|
|
|
|
ss << string_type;
|
|
|
|
}
|
|
|
|
*value = ss.str();
|
|
|
|
return true;
|
|
|
|
}
|
|
|
|
|
|
|
|
bool ParseVectorCompressionType(
|
|
|
|
const std::string& value,
|
|
|
|
std::vector<CompressionType>* compression_per_level) {
|
|
|
|
compression_per_level->clear();
|
|
|
|
size_t start = 0;
|
|
|
|
while (start < value.size()) {
|
|
|
|
size_t end = value.find(':', start);
|
|
|
|
bool is_ok;
|
|
|
|
CompressionType type;
|
|
|
|
if (end == std::string::npos) {
|
2015-11-18 01:41:54 +01:00
|
|
|
is_ok = ParseEnum<CompressionType>(compression_type_string_map,
|
|
|
|
value.substr(start), &type);
|
2015-10-03 00:35:32 +02:00
|
|
|
if (!is_ok) {
|
|
|
|
return false;
|
|
|
|
}
|
|
|
|
compression_per_level->emplace_back(type);
|
|
|
|
break;
|
|
|
|
} else {
|
2015-11-18 01:41:54 +01:00
|
|
|
is_ok = ParseEnum<CompressionType>(
|
|
|
|
compression_type_string_map, value.substr(start, end - start), &type);
|
2015-10-03 00:35:32 +02:00
|
|
|
if (!is_ok) {
|
|
|
|
return false;
|
|
|
|
}
|
|
|
|
compression_per_level->emplace_back(type);
|
|
|
|
start = end + 1;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
return true;
|
|
|
|
}
|
|
|
|
|
2017-10-20 00:19:20 +02:00
|
|
|
// This is to handle backward compatibility, where compaction_options_fifo
|
|
|
|
// could be assigned a single scalar value, say, like "23", which would be
|
|
|
|
// assigned to max_table_files_size.
|
|
|
|
bool FIFOCompactionOptionsSpecialCase(const std::string& opt_str,
|
|
|
|
CompactionOptionsFIFO* options) {
|
|
|
|
if (opt_str.find("=") != std::string::npos) {
|
|
|
|
// New format. Go do your new parsing using ParseStructOptions.
|
|
|
|
return false;
|
|
|
|
}
|
|
|
|
|
|
|
|
// Old format. Parse just a single uint64_t value.
|
|
|
|
options->max_table_files_size = ParseUint64(opt_str);
|
|
|
|
return true;
|
|
|
|
}
|
|
|
|
|
|
|
|
template <typename T>
|
|
|
|
bool SerializeStruct(
|
|
|
|
const T& options, std::string* value,
|
2019-05-15 23:19:04 +02:00
|
|
|
const std::unordered_map<std::string, OptionTypeInfo>& type_info_map) {
|
2017-10-20 00:19:20 +02:00
|
|
|
std::string opt_str;
|
|
|
|
Status s = GetStringFromStruct(&opt_str, options, type_info_map, ";");
|
|
|
|
if (!s.ok()) {
|
|
|
|
return false;
|
|
|
|
}
|
|
|
|
*value = "{" + opt_str + "}";
|
|
|
|
return true;
|
|
|
|
}
|
|
|
|
|
|
|
|
template <typename T>
|
|
|
|
bool ParseSingleStructOption(
|
|
|
|
const std::string& opt_val_str, T* options,
|
2019-05-15 23:19:04 +02:00
|
|
|
const std::unordered_map<std::string, OptionTypeInfo>& type_info_map) {
|
2017-10-20 00:19:20 +02:00
|
|
|
size_t end = opt_val_str.find('=');
|
|
|
|
std::string key = opt_val_str.substr(0, end);
|
|
|
|
std::string value = opt_val_str.substr(end + 1);
|
|
|
|
auto iter = type_info_map.find(key);
|
|
|
|
if (iter == type_info_map.end()) {
|
|
|
|
return false;
|
|
|
|
}
|
|
|
|
const auto& opt_info = iter->second;
|
2019-05-30 19:43:34 +02:00
|
|
|
if (opt_info.verification == OptionVerificationType::kDeprecated) {
|
|
|
|
// Should also skip deprecated sub-options such as
|
|
|
|
// fifo_compaction_options_type_info.ttl
|
|
|
|
return true;
|
|
|
|
}
|
2017-10-20 00:19:20 +02:00
|
|
|
return ParseOptionHelper(
|
|
|
|
reinterpret_cast<char*>(options) + opt_info.mutable_offset, opt_info.type,
|
|
|
|
value);
|
|
|
|
}
|
|
|
|
|
|
|
|
template <typename T>
|
|
|
|
bool ParseStructOptions(
|
|
|
|
const std::string& opt_str, T* options,
|
2019-05-15 23:19:04 +02:00
|
|
|
const std::unordered_map<std::string, OptionTypeInfo>& type_info_map) {
|
2017-10-20 00:19:20 +02:00
|
|
|
assert(!opt_str.empty());
|
|
|
|
|
|
|
|
size_t start = 0;
|
|
|
|
if (opt_str[0] == '{') {
|
|
|
|
start++;
|
|
|
|
}
|
|
|
|
while ((start != std::string::npos) && (start < opt_str.size())) {
|
|
|
|
if (opt_str[start] == '}') {
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
size_t end = opt_str.find(';', start);
|
|
|
|
size_t len = (end == std::string::npos) ? end : end - start;
|
|
|
|
if (!ParseSingleStructOption(opt_str.substr(start, len), options,
|
|
|
|
type_info_map)) {
|
|
|
|
return false;
|
|
|
|
}
|
|
|
|
start = (end == std::string::npos) ? end : end + 1;
|
|
|
|
}
|
|
|
|
return true;
|
|
|
|
}
|
2018-06-27 00:56:26 +02:00
|
|
|
} // anonymouse namespace
|
2017-10-20 00:19:20 +02:00
|
|
|
|
2015-10-03 00:35:32 +02:00
|
|
|
bool ParseSliceTransformHelper(
|
|
|
|
const std::string& kFixedPrefixName, const std::string& kCappedPrefixName,
|
|
|
|
const std::string& value,
|
|
|
|
std::shared_ptr<const SliceTransform>* slice_transform) {
|
2018-05-21 23:33:55 +02:00
|
|
|
const char* no_op_name = "rocksdb.Noop";
|
|
|
|
size_t no_op_length = strlen(no_op_name);
|
2015-10-03 00:35:32 +02:00
|
|
|
auto& pe_value = value;
|
|
|
|
if (pe_value.size() > kFixedPrefixName.size() &&
|
|
|
|
pe_value.compare(0, kFixedPrefixName.size(), kFixedPrefixName) == 0) {
|
|
|
|
int prefix_length = ParseInt(trim(value.substr(kFixedPrefixName.size())));
|
|
|
|
slice_transform->reset(NewFixedPrefixTransform(prefix_length));
|
|
|
|
} else if (pe_value.size() > kCappedPrefixName.size() &&
|
|
|
|
pe_value.compare(0, kCappedPrefixName.size(), kCappedPrefixName) ==
|
|
|
|
0) {
|
|
|
|
int prefix_length =
|
|
|
|
ParseInt(trim(pe_value.substr(kCappedPrefixName.size())));
|
|
|
|
slice_transform->reset(NewCappedPrefixTransform(prefix_length));
|
2018-05-21 23:33:55 +02:00
|
|
|
} else if (pe_value.size() == no_op_length &&
|
|
|
|
pe_value.compare(0, no_op_length, no_op_name) == 0) {
|
|
|
|
const SliceTransform* no_op_transform = NewNoopTransform();
|
|
|
|
slice_transform->reset(no_op_transform);
|
2015-10-11 21:17:42 +02:00
|
|
|
} else if (value == kNullptrString) {
|
2015-10-03 00:35:32 +02:00
|
|
|
slice_transform->reset();
|
|
|
|
} else {
|
|
|
|
return false;
|
|
|
|
}
|
|
|
|
|
|
|
|
return true;
|
|
|
|
}
|
|
|
|
|
|
|
|
bool ParseSliceTransform(
|
|
|
|
const std::string& value,
|
|
|
|
std::shared_ptr<const SliceTransform>* slice_transform) {
|
|
|
|
// While we normally don't convert the string representation of a
|
|
|
|
// pointer-typed option into its instance, here we do so for backward
|
|
|
|
// compatibility as we allow this action in SetOption().
|
|
|
|
|
|
|
|
// TODO(yhchiang): A possible better place for these serialization /
|
|
|
|
// deserialization is inside the class definition of pointer-typed
|
|
|
|
// option itself, but this requires a bigger change of public API.
|
|
|
|
bool result =
|
|
|
|
ParseSliceTransformHelper("fixed:", "capped:", value, slice_transform);
|
|
|
|
if (result) {
|
|
|
|
return result;
|
|
|
|
}
|
|
|
|
result = ParseSliceTransformHelper(
|
|
|
|
"rocksdb.FixedPrefix.", "rocksdb.CappedPrefix.", value, slice_transform);
|
|
|
|
if (result) {
|
|
|
|
return result;
|
|
|
|
}
|
|
|
|
// TODO(yhchiang): we can further support other default
|
|
|
|
// SliceTransforms here.
|
|
|
|
return false;
|
|
|
|
}
|
|
|
|
|
2015-08-27 01:13:56 +02:00
|
|
|
bool ParseOptionHelper(char* opt_address, const OptionType& opt_type,
|
|
|
|
const std::string& value) {
|
|
|
|
switch (opt_type) {
|
|
|
|
case OptionType::kBoolean:
|
|
|
|
*reinterpret_cast<bool*>(opt_address) = ParseBoolean("", value);
|
|
|
|
break;
|
|
|
|
case OptionType::kInt:
|
|
|
|
*reinterpret_cast<int*>(opt_address) = ParseInt(value);
|
|
|
|
break;
|
2019-03-12 21:46:12 +01:00
|
|
|
case OptionType::kInt32T:
|
|
|
|
*reinterpret_cast<int32_t*>(opt_address) = ParseInt32(value);
|
|
|
|
break;
|
|
|
|
case OptionType::kInt64T:
|
|
|
|
PutUnaligned(reinterpret_cast<int64_t*>(opt_address), ParseInt64(value));
|
|
|
|
break;
|
2016-09-14 06:12:43 +02:00
|
|
|
case OptionType::kVectorInt:
|
|
|
|
*reinterpret_cast<std::vector<int>*>(opt_address) = ParseVectorInt(value);
|
|
|
|
break;
|
2015-08-27 01:13:56 +02:00
|
|
|
case OptionType::kUInt:
|
|
|
|
*reinterpret_cast<unsigned int*>(opt_address) = ParseUint32(value);
|
|
|
|
break;
|
|
|
|
case OptionType::kUInt32T:
|
|
|
|
*reinterpret_cast<uint32_t*>(opt_address) = ParseUint32(value);
|
|
|
|
break;
|
|
|
|
case OptionType::kUInt64T:
|
2017-04-22 05:41:37 +02:00
|
|
|
PutUnaligned(reinterpret_cast<uint64_t*>(opt_address), ParseUint64(value));
|
2015-08-27 01:13:56 +02:00
|
|
|
break;
|
|
|
|
case OptionType::kSizeT:
|
2017-04-22 05:41:37 +02:00
|
|
|
PutUnaligned(reinterpret_cast<size_t*>(opt_address), ParseSizeT(value));
|
2015-08-27 01:13:56 +02:00
|
|
|
break;
|
|
|
|
case OptionType::kString:
|
|
|
|
*reinterpret_cast<std::string*>(opt_address) = value;
|
|
|
|
break;
|
|
|
|
case OptionType::kDouble:
|
|
|
|
*reinterpret_cast<double*>(opt_address) = ParseDouble(value);
|
|
|
|
break;
|
|
|
|
case OptionType::kCompactionStyle:
|
2015-11-18 01:41:54 +01:00
|
|
|
return ParseEnum<CompactionStyle>(
|
|
|
|
compaction_style_string_map, value,
|
|
|
|
reinterpret_cast<CompactionStyle*>(opt_address));
|
2017-03-02 19:08:49 +01:00
|
|
|
case OptionType::kCompactionPri:
|
|
|
|
return ParseEnum<CompactionPri>(
|
|
|
|
compaction_pri_string_map, value,
|
|
|
|
reinterpret_cast<CompactionPri*>(opt_address));
|
2015-10-03 00:35:32 +02:00
|
|
|
case OptionType::kCompressionType:
|
2015-11-18 01:41:54 +01:00
|
|
|
return ParseEnum<CompressionType>(
|
|
|
|
compression_type_string_map, value,
|
|
|
|
reinterpret_cast<CompressionType*>(opt_address));
|
2015-10-03 00:35:32 +02:00
|
|
|
case OptionType::kVectorCompressionType:
|
|
|
|
return ParseVectorCompressionType(
|
|
|
|
value, reinterpret_cast<std::vector<CompressionType>*>(opt_address));
|
|
|
|
case OptionType::kSliceTransform:
|
|
|
|
return ParseSliceTransform(
|
|
|
|
value, reinterpret_cast<std::shared_ptr<const SliceTransform>*>(
|
|
|
|
opt_address));
|
2015-10-11 21:17:42 +02:00
|
|
|
case OptionType::kChecksumType:
|
2015-11-18 01:41:54 +01:00
|
|
|
return ParseEnum<ChecksumType>(
|
|
|
|
checksum_type_string_map, value,
|
|
|
|
reinterpret_cast<ChecksumType*>(opt_address));
|
2015-10-11 21:17:42 +02:00
|
|
|
case OptionType::kBlockBasedTableIndexType:
|
2015-10-30 23:58:46 +01:00
|
|
|
return ParseEnum<BlockBasedTableOptions::IndexType>(
|
2015-11-18 01:41:54 +01:00
|
|
|
block_base_table_index_type_string_map, value,
|
|
|
|
reinterpret_cast<BlockBasedTableOptions::IndexType*>(opt_address));
|
2018-07-28 00:35:41 +02:00
|
|
|
case OptionType::kBlockBasedTableDataBlockIndexType:
|
|
|
|
return ParseEnum<BlockBasedTableOptions::DataBlockIndexType>(
|
|
|
|
block_base_table_data_block_index_type_string_map, value,
|
|
|
|
reinterpret_cast<BlockBasedTableOptions::DataBlockIndexType*>(
|
|
|
|
opt_address));
|
2019-04-22 17:17:45 +02:00
|
|
|
case OptionType::kBlockBasedTableIndexShorteningMode:
|
|
|
|
return ParseEnum<BlockBasedTableOptions::IndexShorteningMode>(
|
2019-05-04 02:26:20 +02:00
|
|
|
block_base_table_index_shortening_mode_string_map, value,
|
|
|
|
reinterpret_cast<BlockBasedTableOptions::IndexShorteningMode*>(
|
|
|
|
opt_address));
|
2015-10-28 18:46:01 +01:00
|
|
|
case OptionType::kEncodingType:
|
2015-11-18 01:41:54 +01:00
|
|
|
return ParseEnum<EncodingType>(
|
|
|
|
encoding_type_string_map, value,
|
|
|
|
reinterpret_cast<EncodingType*>(opt_address));
|
2016-03-02 19:34:14 +01:00
|
|
|
case OptionType::kWALRecoveryMode:
|
|
|
|
return ParseEnum<WALRecoveryMode>(
|
|
|
|
wal_recovery_mode_string_map, value,
|
|
|
|
reinterpret_cast<WALRecoveryMode*>(opt_address));
|
|
|
|
case OptionType::kAccessHint:
|
|
|
|
return ParseEnum<DBOptions::AccessHint>(
|
|
|
|
access_hint_string_map, value,
|
|
|
|
reinterpret_cast<DBOptions::AccessHint*>(opt_address));
|
|
|
|
case OptionType::kInfoLogLevel:
|
|
|
|
return ParseEnum<InfoLogLevel>(
|
|
|
|
info_log_level_string_map, value,
|
|
|
|
reinterpret_cast<InfoLogLevel*>(opt_address));
|
2017-10-20 00:19:20 +02:00
|
|
|
case OptionType::kCompactionOptionsFIFO: {
|
|
|
|
if (!FIFOCompactionOptionsSpecialCase(
|
|
|
|
value, reinterpret_cast<CompactionOptionsFIFO*>(opt_address))) {
|
|
|
|
return ParseStructOptions<CompactionOptionsFIFO>(
|
|
|
|
value, reinterpret_cast<CompactionOptionsFIFO*>(opt_address),
|
|
|
|
fifo_compaction_options_type_info);
|
|
|
|
}
|
|
|
|
return true;
|
|
|
|
}
|
2017-11-28 19:35:17 +01:00
|
|
|
case OptionType::kLRUCacheOptions: {
|
|
|
|
return ParseStructOptions<LRUCacheOptions>(value,
|
|
|
|
reinterpret_cast<LRUCacheOptions*>(opt_address),
|
|
|
|
lru_cache_options_type_info);
|
|
|
|
}
|
2017-12-11 22:12:12 +01:00
|
|
|
case OptionType::kCompactionOptionsUniversal:
|
|
|
|
return ParseStructOptions<CompactionOptionsUniversal>(
|
|
|
|
value, reinterpret_cast<CompactionOptionsUniversal*>(opt_address),
|
|
|
|
universal_compaction_options_type_info);
|
|
|
|
case OptionType::kCompactionStopStyle:
|
|
|
|
return ParseEnum<CompactionStopStyle>(
|
|
|
|
compaction_stop_style_string_map, value,
|
|
|
|
reinterpret_cast<CompactionStopStyle*>(opt_address));
|
2015-08-27 01:13:56 +02:00
|
|
|
default:
|
|
|
|
return false;
|
|
|
|
}
|
|
|
|
return true;
|
|
|
|
}
|
|
|
|
|
|
|
|
bool SerializeSingleOptionHelper(const char* opt_address,
|
|
|
|
const OptionType opt_type,
|
|
|
|
std::string* value) {
|
2016-10-14 22:09:18 +02:00
|
|
|
|
2015-08-27 01:13:56 +02:00
|
|
|
assert(value);
|
|
|
|
switch (opt_type) {
|
|
|
|
case OptionType::kBoolean:
|
|
|
|
*value = *(reinterpret_cast<const bool*>(opt_address)) ? "true" : "false";
|
|
|
|
break;
|
|
|
|
case OptionType::kInt:
|
|
|
|
*value = ToString(*(reinterpret_cast<const int*>(opt_address)));
|
|
|
|
break;
|
2019-03-12 21:46:12 +01:00
|
|
|
case OptionType::kInt32T:
|
|
|
|
*value = ToString(*(reinterpret_cast<const int32_t*>(opt_address)));
|
|
|
|
break;
|
|
|
|
case OptionType::kInt64T:
|
|
|
|
{
|
|
|
|
int64_t v;
|
|
|
|
GetUnaligned(reinterpret_cast<const int64_t*>(opt_address), &v);
|
|
|
|
*value = ToString(v);
|
|
|
|
}
|
|
|
|
break;
|
2016-09-14 06:12:43 +02:00
|
|
|
case OptionType::kVectorInt:
|
|
|
|
return SerializeIntVector(
|
|
|
|
*reinterpret_cast<const std::vector<int>*>(opt_address), value);
|
2015-08-27 01:13:56 +02:00
|
|
|
case OptionType::kUInt:
|
|
|
|
*value = ToString(*(reinterpret_cast<const unsigned int*>(opt_address)));
|
|
|
|
break;
|
|
|
|
case OptionType::kUInt32T:
|
|
|
|
*value = ToString(*(reinterpret_cast<const uint32_t*>(opt_address)));
|
|
|
|
break;
|
|
|
|
case OptionType::kUInt64T:
|
2017-04-22 05:41:37 +02:00
|
|
|
{
|
|
|
|
uint64_t v;
|
|
|
|
GetUnaligned(reinterpret_cast<const uint64_t*>(opt_address), &v);
|
|
|
|
*value = ToString(v);
|
|
|
|
}
|
2015-08-27 01:13:56 +02:00
|
|
|
break;
|
|
|
|
case OptionType::kSizeT:
|
2017-04-22 05:41:37 +02:00
|
|
|
{
|
|
|
|
size_t v;
|
|
|
|
GetUnaligned(reinterpret_cast<const size_t*>(opt_address), &v);
|
|
|
|
*value = ToString(v);
|
|
|
|
}
|
2015-08-27 01:13:56 +02:00
|
|
|
break;
|
|
|
|
case OptionType::kDouble:
|
|
|
|
*value = ToString(*(reinterpret_cast<const double*>(opt_address)));
|
|
|
|
break;
|
|
|
|
case OptionType::kString:
|
RocksDB Options file format and its serialization / deserialization.
Summary:
This patch defines the format of RocksDB options file, which
follows the INI file format, and implements functions for its
serialization and deserialization. An example RocksDB options
file can be found in examples/rocksdb_option_file_example.ini.
A typical RocksDB options file has three sections, which are
Version, DBOptions, and more than one CFOptions. The RocksDB
options file in general follows the basic INI file format
with the following extensions / modifications:
* Escaped characters
We escaped the following characters:
- \n -- line feed - new line
- \r -- carriage return
- \\ -- backslash \
- \: -- colon symbol :
- \# -- hash tag #
* Comments
We support # style comments. Comments can appear at the ending
part of a line.
* Statements
A statement is of the form option_name = value.
Each statement contains a '=', where extra white-spaces
are supported. However, we don't support multi-lined statement.
Furthermore, each line can only contain at most one statement.
* Section
Sections are of the form [SecitonTitle "SectionArgument"],
where section argument is optional.
* List
We use colon-separated string to represent a list.
For instance, n1:n2:n3:n4 is a list containing four values.
Below is an example of a RocksDB options file:
[Version]
rocksdb_version=4.0.0
options_file_version=1.0
[DBOptions]
max_open_files=12345
max_background_flushes=301
[CFOptions "default"]
[CFOptions "the second column family"]
[CFOptions "the third column family"]
Test Plan: Added many tests in options_test.cc
Reviewers: igor, IslamAbdelRahman, sdong, anthony
Reviewed By: anthony
Subscribers: maykov, dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D46059
2015-09-29 23:42:40 +02:00
|
|
|
*value = EscapeOptionString(
|
|
|
|
*(reinterpret_cast<const std::string*>(opt_address)));
|
2015-08-27 01:13:56 +02:00
|
|
|
break;
|
|
|
|
case OptionType::kCompactionStyle:
|
2015-11-18 01:41:54 +01:00
|
|
|
return SerializeEnum<CompactionStyle>(
|
|
|
|
compaction_style_string_map,
|
2015-10-30 23:58:46 +01:00
|
|
|
*(reinterpret_cast<const CompactionStyle*>(opt_address)), value);
|
2017-03-02 19:08:49 +01:00
|
|
|
case OptionType::kCompactionPri:
|
|
|
|
return SerializeEnum<CompactionPri>(
|
|
|
|
compaction_pri_string_map,
|
|
|
|
*(reinterpret_cast<const CompactionPri*>(opt_address)), value);
|
2015-10-03 00:35:32 +02:00
|
|
|
case OptionType::kCompressionType:
|
2015-11-18 01:41:54 +01:00
|
|
|
return SerializeEnum<CompressionType>(
|
|
|
|
compression_type_string_map,
|
2015-10-03 00:35:32 +02:00
|
|
|
*(reinterpret_cast<const CompressionType*>(opt_address)), value);
|
|
|
|
case OptionType::kVectorCompressionType:
|
|
|
|
return SerializeVectorCompressionType(
|
|
|
|
*(reinterpret_cast<const std::vector<CompressionType>*>(opt_address)),
|
|
|
|
value);
|
|
|
|
break;
|
|
|
|
case OptionType::kSliceTransform: {
|
|
|
|
const auto* slice_transform_ptr =
|
|
|
|
reinterpret_cast<const std::shared_ptr<const SliceTransform>*>(
|
|
|
|
opt_address);
|
|
|
|
*value = slice_transform_ptr->get() ? slice_transform_ptr->get()->Name()
|
2015-10-11 21:17:42 +02:00
|
|
|
: kNullptrString;
|
2015-10-03 00:35:32 +02:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
case OptionType::kTableFactory: {
|
|
|
|
const auto* table_factory_ptr =
|
|
|
|
reinterpret_cast<const std::shared_ptr<const TableFactory>*>(
|
|
|
|
opt_address);
|
|
|
|
*value = table_factory_ptr->get() ? table_factory_ptr->get()->Name()
|
2015-10-11 21:17:42 +02:00
|
|
|
: kNullptrString;
|
2015-10-03 00:35:32 +02:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
case OptionType::kComparator: {
|
|
|
|
// it's a const pointer of const Comparator*
|
|
|
|
const auto* ptr = reinterpret_cast<const Comparator* const*>(opt_address);
|
Add OptionsUtil::LoadOptionsFromFile() API
Summary:
This patch adds OptionsUtil::LoadOptionsFromFile() and
OptionsUtil::LoadLatestOptionsFromDB(), which allow developers
to construct DBOptions and ColumnFamilyOptions from a RocksDB
options file. Note that most pointer-typed options such as
merge_operator will not be constructed.
With this API, developers no longer need to remember all the
options in order to reopen an existing rocksdb instance like
the following:
DBOptions db_options;
std::vector<std::string> cf_names;
std::vector<ColumnFamilyOptions> cf_opts;
// Load primitive-typed options from an existing DB
OptionsUtil::LoadLatestOptionsFromDB(
dbname, &db_options, &cf_names, &cf_opts);
// Initialize necessary pointer-typed options
cf_opts[0].merge_operator.reset(new MyMergeOperator());
...
// Construct the vector of ColumnFamilyDescriptor
std::vector<ColumnFamilyDescriptor> cf_descs;
for (size_t i = 0; i < cf_opts.size(); ++i) {
cf_descs.emplace_back(cf_names[i], cf_opts[i]);
}
// Open the DB
DB* db = nullptr;
std::vector<ColumnFamilyHandle*> cf_handles;
auto s = DB::Open(db_options, dbname, cf_descs,
&handles, &db);
Test Plan:
Augment existing tests in column_family_test
options_test
db_test
Reviewers: igor, IslamAbdelRahman, sdong, anthony
Reviewed By: anthony
Subscribers: dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D49095
2015-11-12 15:52:43 +01:00
|
|
|
// Since the user-specified comparator will be wrapped by
|
|
|
|
// InternalKeyComparator, we should persist the user-specified one
|
|
|
|
// instead of InternalKeyComparator.
|
2017-07-29 01:23:50 +02:00
|
|
|
if (*ptr == nullptr) {
|
|
|
|
*value = kNullptrString;
|
Add OptionsUtil::LoadOptionsFromFile() API
Summary:
This patch adds OptionsUtil::LoadOptionsFromFile() and
OptionsUtil::LoadLatestOptionsFromDB(), which allow developers
to construct DBOptions and ColumnFamilyOptions from a RocksDB
options file. Note that most pointer-typed options such as
merge_operator will not be constructed.
With this API, developers no longer need to remember all the
options in order to reopen an existing rocksdb instance like
the following:
DBOptions db_options;
std::vector<std::string> cf_names;
std::vector<ColumnFamilyOptions> cf_opts;
// Load primitive-typed options from an existing DB
OptionsUtil::LoadLatestOptionsFromDB(
dbname, &db_options, &cf_names, &cf_opts);
// Initialize necessary pointer-typed options
cf_opts[0].merge_operator.reset(new MyMergeOperator());
...
// Construct the vector of ColumnFamilyDescriptor
std::vector<ColumnFamilyDescriptor> cf_descs;
for (size_t i = 0; i < cf_opts.size(); ++i) {
cf_descs.emplace_back(cf_names[i], cf_opts[i]);
}
// Open the DB
DB* db = nullptr;
std::vector<ColumnFamilyHandle*> cf_handles;
auto s = DB::Open(db_options, dbname, cf_descs,
&handles, &db);
Test Plan:
Augment existing tests in column_family_test
options_test
db_test
Reviewers: igor, IslamAbdelRahman, sdong, anthony
Reviewed By: anthony
Subscribers: dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D49095
2015-11-12 15:52:43 +01:00
|
|
|
} else {
|
2017-07-29 01:23:50 +02:00
|
|
|
const Comparator* root_comp = (*ptr)->GetRootComparator();
|
|
|
|
if (root_comp == nullptr) {
|
|
|
|
root_comp = (*ptr);
|
|
|
|
}
|
|
|
|
*value = root_comp->Name();
|
Add OptionsUtil::LoadOptionsFromFile() API
Summary:
This patch adds OptionsUtil::LoadOptionsFromFile() and
OptionsUtil::LoadLatestOptionsFromDB(), which allow developers
to construct DBOptions and ColumnFamilyOptions from a RocksDB
options file. Note that most pointer-typed options such as
merge_operator will not be constructed.
With this API, developers no longer need to remember all the
options in order to reopen an existing rocksdb instance like
the following:
DBOptions db_options;
std::vector<std::string> cf_names;
std::vector<ColumnFamilyOptions> cf_opts;
// Load primitive-typed options from an existing DB
OptionsUtil::LoadLatestOptionsFromDB(
dbname, &db_options, &cf_names, &cf_opts);
// Initialize necessary pointer-typed options
cf_opts[0].merge_operator.reset(new MyMergeOperator());
...
// Construct the vector of ColumnFamilyDescriptor
std::vector<ColumnFamilyDescriptor> cf_descs;
for (size_t i = 0; i < cf_opts.size(); ++i) {
cf_descs.emplace_back(cf_names[i], cf_opts[i]);
}
// Open the DB
DB* db = nullptr;
std::vector<ColumnFamilyHandle*> cf_handles;
auto s = DB::Open(db_options, dbname, cf_descs,
&handles, &db);
Test Plan:
Augment existing tests in column_family_test
options_test
db_test
Reviewers: igor, IslamAbdelRahman, sdong, anthony
Reviewed By: anthony
Subscribers: dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D49095
2015-11-12 15:52:43 +01:00
|
|
|
}
|
2015-10-03 00:35:32 +02:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
case OptionType::kCompactionFilter: {
|
|
|
|
// it's a const pointer of const CompactionFilter*
|
|
|
|
const auto* ptr =
|
|
|
|
reinterpret_cast<const CompactionFilter* const*>(opt_address);
|
2015-10-11 21:17:42 +02:00
|
|
|
*value = *ptr ? (*ptr)->Name() : kNullptrString;
|
2015-10-03 00:35:32 +02:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
case OptionType::kCompactionFilterFactory: {
|
|
|
|
const auto* ptr =
|
|
|
|
reinterpret_cast<const std::shared_ptr<CompactionFilterFactory>*>(
|
|
|
|
opt_address);
|
2015-10-11 21:17:42 +02:00
|
|
|
*value = ptr->get() ? ptr->get()->Name() : kNullptrString;
|
2015-10-03 00:35:32 +02:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
case OptionType::kMemTableRepFactory: {
|
|
|
|
const auto* ptr =
|
|
|
|
reinterpret_cast<const std::shared_ptr<MemTableRepFactory>*>(
|
|
|
|
opt_address);
|
2015-10-11 21:17:42 +02:00
|
|
|
*value = ptr->get() ? ptr->get()->Name() : kNullptrString;
|
2015-10-03 00:35:32 +02:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
case OptionType::kMergeOperator: {
|
|
|
|
const auto* ptr =
|
|
|
|
reinterpret_cast<const std::shared_ptr<MergeOperator>*>(opt_address);
|
2015-10-11 21:17:42 +02:00
|
|
|
*value = ptr->get() ? ptr->get()->Name() : kNullptrString;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
case OptionType::kFilterPolicy: {
|
|
|
|
const auto* ptr =
|
|
|
|
reinterpret_cast<const std::shared_ptr<FilterPolicy>*>(opt_address);
|
|
|
|
*value = ptr->get() ? ptr->get()->Name() : kNullptrString;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
case OptionType::kChecksumType:
|
2015-11-18 01:41:54 +01:00
|
|
|
return SerializeEnum<ChecksumType>(
|
|
|
|
checksum_type_string_map,
|
2015-10-11 21:17:42 +02:00
|
|
|
*reinterpret_cast<const ChecksumType*>(opt_address), value);
|
|
|
|
case OptionType::kBlockBasedTableIndexType:
|
2015-10-30 23:58:46 +01:00
|
|
|
return SerializeEnum<BlockBasedTableOptions::IndexType>(
|
|
|
|
block_base_table_index_type_string_map,
|
2015-10-11 21:17:42 +02:00
|
|
|
*reinterpret_cast<const BlockBasedTableOptions::IndexType*>(
|
|
|
|
opt_address),
|
|
|
|
value);
|
2018-07-28 00:35:41 +02:00
|
|
|
case OptionType::kBlockBasedTableDataBlockIndexType:
|
|
|
|
return SerializeEnum<BlockBasedTableOptions::DataBlockIndexType>(
|
|
|
|
block_base_table_data_block_index_type_string_map,
|
|
|
|
*reinterpret_cast<const BlockBasedTableOptions::DataBlockIndexType*>(
|
|
|
|
opt_address),
|
|
|
|
value);
|
2019-04-22 17:17:45 +02:00
|
|
|
case OptionType::kBlockBasedTableIndexShorteningMode:
|
|
|
|
return SerializeEnum<BlockBasedTableOptions::IndexShorteningMode>(
|
|
|
|
block_base_table_index_shortening_mode_string_map,
|
|
|
|
*reinterpret_cast<const BlockBasedTableOptions::IndexShorteningMode*>(
|
|
|
|
opt_address),
|
|
|
|
value);
|
2015-10-11 21:17:42 +02:00
|
|
|
case OptionType::kFlushBlockPolicyFactory: {
|
|
|
|
const auto* ptr =
|
|
|
|
reinterpret_cast<const std::shared_ptr<FlushBlockPolicyFactory>*>(
|
|
|
|
opt_address);
|
|
|
|
*value = ptr->get() ? ptr->get()->Name() : kNullptrString;
|
2015-10-03 00:35:32 +02:00
|
|
|
break;
|
|
|
|
}
|
2015-10-28 18:46:01 +01:00
|
|
|
case OptionType::kEncodingType:
|
2015-11-18 01:41:54 +01:00
|
|
|
return SerializeEnum<EncodingType>(
|
|
|
|
encoding_type_string_map,
|
2015-10-28 18:46:01 +01:00
|
|
|
*reinterpret_cast<const EncodingType*>(opt_address), value);
|
2016-03-02 19:34:14 +01:00
|
|
|
case OptionType::kWALRecoveryMode:
|
|
|
|
return SerializeEnum<WALRecoveryMode>(
|
|
|
|
wal_recovery_mode_string_map,
|
|
|
|
*reinterpret_cast<const WALRecoveryMode*>(opt_address), value);
|
|
|
|
case OptionType::kAccessHint:
|
|
|
|
return SerializeEnum<DBOptions::AccessHint>(
|
|
|
|
access_hint_string_map,
|
|
|
|
*reinterpret_cast<const DBOptions::AccessHint*>(opt_address), value);
|
|
|
|
case OptionType::kInfoLogLevel:
|
|
|
|
return SerializeEnum<InfoLogLevel>(
|
|
|
|
info_log_level_string_map,
|
|
|
|
*reinterpret_cast<const InfoLogLevel*>(opt_address), value);
|
2017-12-11 22:12:12 +01:00
|
|
|
case OptionType::kCompactionOptionsFIFO:
|
2017-10-20 00:19:20 +02:00
|
|
|
return SerializeStruct<CompactionOptionsFIFO>(
|
|
|
|
*reinterpret_cast<const CompactionOptionsFIFO*>(opt_address), value,
|
|
|
|
fifo_compaction_options_type_info);
|
2017-12-11 22:12:12 +01:00
|
|
|
case OptionType::kCompactionOptionsUniversal:
|
|
|
|
return SerializeStruct<CompactionOptionsUniversal>(
|
|
|
|
*reinterpret_cast<const CompactionOptionsUniversal*>(opt_address),
|
|
|
|
value, universal_compaction_options_type_info);
|
|
|
|
case OptionType::kCompactionStopStyle:
|
|
|
|
return SerializeEnum<CompactionStopStyle>(
|
|
|
|
compaction_stop_style_string_map,
|
|
|
|
*reinterpret_cast<const CompactionStopStyle*>(opt_address), value);
|
2015-08-27 01:13:56 +02:00
|
|
|
default:
|
|
|
|
return false;
|
|
|
|
}
|
|
|
|
return true;
|
|
|
|
}
|
|
|
|
|
2014-11-05 01:23:05 +01:00
|
|
|
Status GetMutableOptionsFromStrings(
|
2014-09-17 21:49:13 +02:00
|
|
|
const MutableCFOptions& base_options,
|
|
|
|
const std::unordered_map<std::string, std::string>& options_map,
|
2018-04-11 03:48:49 +02:00
|
|
|
Logger* info_log, MutableCFOptions* new_options) {
|
2014-09-17 21:49:13 +02:00
|
|
|
assert(new_options);
|
|
|
|
*new_options = base_options;
|
2014-12-22 22:18:57 +01:00
|
|
|
for (const auto& o : options_map) {
|
|
|
|
try {
|
2016-09-14 06:12:43 +02:00
|
|
|
auto iter = cf_options_type_info.find(o.first);
|
|
|
|
if (iter == cf_options_type_info.end()) {
|
|
|
|
return Status::InvalidArgument("Unrecognized option: " + o.first);
|
|
|
|
}
|
|
|
|
const auto& opt_info = iter->second;
|
|
|
|
if (!opt_info.is_mutable) {
|
|
|
|
return Status::InvalidArgument("Option not changeable: " + o.first);
|
|
|
|
}
|
2018-04-11 03:48:49 +02:00
|
|
|
if (opt_info.verification == OptionVerificationType::kDeprecated) {
|
|
|
|
// log warning when user tries to set a deprecated option but don't fail
|
|
|
|
// the call for compatibility.
|
|
|
|
ROCKS_LOG_WARN(info_log, "%s is a deprecated option and cannot be set",
|
|
|
|
o.first.c_str());
|
|
|
|
continue;
|
|
|
|
}
|
2016-09-14 06:12:43 +02:00
|
|
|
bool is_ok = ParseOptionHelper(
|
|
|
|
reinterpret_cast<char*>(new_options) + opt_info.mutable_offset,
|
|
|
|
opt_info.type, o.second);
|
|
|
|
if (!is_ok) {
|
|
|
|
return Status::InvalidArgument("Error parsing " + o.first);
|
2014-09-17 21:49:13 +02:00
|
|
|
}
|
2015-07-08 16:35:56 +02:00
|
|
|
} catch (std::exception& e) {
|
2016-09-14 06:12:43 +02:00
|
|
|
return Status::InvalidArgument("Error parsing " + o.first + ":" +
|
2014-12-22 22:18:57 +01:00
|
|
|
std::string(e.what()));
|
2014-09-17 21:49:13 +02:00
|
|
|
}
|
|
|
|
}
|
2014-11-05 01:23:05 +01:00
|
|
|
return Status::OK();
|
2014-09-17 21:49:13 +02:00
|
|
|
}
|
|
|
|
|
2016-10-14 21:25:39 +02:00
|
|
|
Status GetMutableDBOptionsFromStrings(
|
|
|
|
const MutableDBOptions& base_options,
|
|
|
|
const std::unordered_map<std::string, std::string>& options_map,
|
|
|
|
MutableDBOptions* new_options) {
|
|
|
|
assert(new_options);
|
|
|
|
*new_options = base_options;
|
|
|
|
for (const auto& o : options_map) {
|
|
|
|
try {
|
|
|
|
auto iter = db_options_type_info.find(o.first);
|
|
|
|
if (iter == db_options_type_info.end()) {
|
|
|
|
return Status::InvalidArgument("Unrecognized option: " + o.first);
|
|
|
|
}
|
|
|
|
const auto& opt_info = iter->second;
|
|
|
|
if (!opt_info.is_mutable) {
|
|
|
|
return Status::InvalidArgument("Option not changeable: " + o.first);
|
|
|
|
}
|
|
|
|
bool is_ok = ParseOptionHelper(
|
|
|
|
reinterpret_cast<char*>(new_options) + opt_info.mutable_offset,
|
|
|
|
opt_info.type, o.second);
|
|
|
|
if (!is_ok) {
|
|
|
|
return Status::InvalidArgument("Error parsing " + o.first);
|
|
|
|
}
|
|
|
|
} catch (std::exception& e) {
|
|
|
|
return Status::InvalidArgument("Error parsing " + o.first + ":" +
|
|
|
|
std::string(e.what()));
|
|
|
|
}
|
|
|
|
}
|
|
|
|
return Status::OK();
|
|
|
|
}
|
|
|
|
|
2014-12-22 22:18:57 +01:00
|
|
|
Status StringToMap(const std::string& opts_str,
|
|
|
|
std::unordered_map<std::string, std::string>* opts_map) {
|
2014-10-10 19:00:12 +02:00
|
|
|
assert(opts_map);
|
|
|
|
// Example:
|
2014-12-22 22:18:57 +01:00
|
|
|
// opts_str = "write_buffer_size=1024;max_write_buffer_number=2;"
|
|
|
|
// "nested_opt={opt1=1;opt2=2};max_bytes_for_level_base=100"
|
2014-10-10 19:00:12 +02:00
|
|
|
size_t pos = 0;
|
|
|
|
std::string opts = trim(opts_str);
|
|
|
|
while (pos < opts.size()) {
|
|
|
|
size_t eq_pos = opts.find('=', pos);
|
|
|
|
if (eq_pos == std::string::npos) {
|
2014-12-22 22:18:57 +01:00
|
|
|
return Status::InvalidArgument("Mismatched key value pair, '=' expected");
|
2014-10-10 19:00:12 +02:00
|
|
|
}
|
|
|
|
std::string key = trim(opts.substr(pos, eq_pos - pos));
|
2014-12-22 22:18:57 +01:00
|
|
|
if (key.empty()) {
|
|
|
|
return Status::InvalidArgument("Empty key found");
|
|
|
|
}
|
2014-10-10 19:00:12 +02:00
|
|
|
|
2014-12-22 22:18:57 +01:00
|
|
|
// skip space after '=' and look for '{' for possible nested options
|
|
|
|
pos = eq_pos + 1;
|
|
|
|
while (pos < opts.size() && isspace(opts[pos])) {
|
|
|
|
++pos;
|
|
|
|
}
|
|
|
|
// Empty value at the end
|
|
|
|
if (pos >= opts.size()) {
|
|
|
|
(*opts_map)[key] = "";
|
2014-10-10 19:00:12 +02:00
|
|
|
break;
|
2014-12-22 22:18:57 +01:00
|
|
|
}
|
|
|
|
if (opts[pos] == '{') {
|
|
|
|
int count = 1;
|
|
|
|
size_t brace_pos = pos + 1;
|
|
|
|
while (brace_pos < opts.size()) {
|
|
|
|
if (opts[brace_pos] == '{') {
|
|
|
|
++count;
|
|
|
|
} else if (opts[brace_pos] == '}') {
|
|
|
|
--count;
|
|
|
|
if (count == 0) {
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
++brace_pos;
|
|
|
|
}
|
|
|
|
// found the matching closing brace
|
|
|
|
if (count == 0) {
|
|
|
|
(*opts_map)[key] = trim(opts.substr(pos + 1, brace_pos - pos - 1));
|
|
|
|
// skip all whitespace and move to the next ';'
|
|
|
|
// brace_pos points to the next position after the matching '}'
|
|
|
|
pos = brace_pos + 1;
|
|
|
|
while (pos < opts.size() && isspace(opts[pos])) {
|
|
|
|
++pos;
|
|
|
|
}
|
|
|
|
if (pos < opts.size() && opts[pos] != ';') {
|
|
|
|
return Status::InvalidArgument(
|
|
|
|
"Unexpected chars after nested options");
|
|
|
|
}
|
|
|
|
++pos;
|
|
|
|
} else {
|
|
|
|
return Status::InvalidArgument(
|
|
|
|
"Mismatched curly braces for nested options");
|
|
|
|
}
|
2014-10-10 19:00:12 +02:00
|
|
|
} else {
|
2014-12-22 22:18:57 +01:00
|
|
|
size_t sc_pos = opts.find(';', pos);
|
|
|
|
if (sc_pos == std::string::npos) {
|
|
|
|
(*opts_map)[key] = trim(opts.substr(pos));
|
|
|
|
// It either ends with a trailing semi-colon or the last key-value pair
|
|
|
|
break;
|
|
|
|
} else {
|
|
|
|
(*opts_map)[key] = trim(opts.substr(pos, sc_pos - pos));
|
|
|
|
}
|
|
|
|
pos = sc_pos + 1;
|
2014-10-10 19:00:12 +02:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2014-12-22 22:18:57 +01:00
|
|
|
return Status::OK();
|
2014-10-10 19:00:12 +02:00
|
|
|
}
|
|
|
|
|
2018-06-28 02:34:07 +02:00
|
|
|
Status ParseCompressionOptions(const std::string& value, const std::string& name,
|
|
|
|
CompressionOptions& compression_opts) {
|
|
|
|
size_t start = 0;
|
|
|
|
size_t end = value.find(':');
|
|
|
|
if (end == std::string::npos) {
|
|
|
|
return Status::InvalidArgument("unable to parse the specified CF option " +
|
|
|
|
name);
|
|
|
|
}
|
|
|
|
compression_opts.window_bits = ParseInt(value.substr(start, end - start));
|
|
|
|
start = end + 1;
|
|
|
|
end = value.find(':', start);
|
|
|
|
if (end == std::string::npos) {
|
|
|
|
return Status::InvalidArgument("unable to parse the specified CF option " +
|
|
|
|
name);
|
|
|
|
}
|
|
|
|
compression_opts.level = ParseInt(value.substr(start, end - start));
|
|
|
|
start = end + 1;
|
|
|
|
if (start >= value.size()) {
|
|
|
|
return Status::InvalidArgument("unable to parse the specified CF option " +
|
|
|
|
name);
|
|
|
|
}
|
|
|
|
end = value.find(':', start);
|
|
|
|
compression_opts.strategy =
|
|
|
|
ParseInt(value.substr(start, value.size() - start));
|
|
|
|
// max_dict_bytes is optional for backwards compatibility
|
|
|
|
if (end != std::string::npos) {
|
|
|
|
start = end + 1;
|
|
|
|
if (start >= value.size()) {
|
|
|
|
return Status::InvalidArgument(
|
|
|
|
"unable to parse the specified CF option " + name);
|
|
|
|
}
|
|
|
|
compression_opts.max_dict_bytes =
|
|
|
|
ParseInt(value.substr(start, value.size() - start));
|
|
|
|
end = value.find(':', start);
|
|
|
|
}
|
|
|
|
// zstd_max_train_bytes is optional for backwards compatibility
|
|
|
|
if (end != std::string::npos) {
|
|
|
|
start = end + 1;
|
|
|
|
if (start >= value.size()) {
|
|
|
|
return Status::InvalidArgument(
|
|
|
|
"unable to parse the specified CF option " + name);
|
|
|
|
}
|
|
|
|
compression_opts.zstd_max_train_bytes =
|
|
|
|
ParseInt(value.substr(start, value.size() - start));
|
|
|
|
end = value.find(':', start);
|
|
|
|
}
|
|
|
|
// enabled is optional for backwards compatibility
|
|
|
|
if (end != std::string::npos) {
|
|
|
|
start = end + 1;
|
|
|
|
if (start >= value.size()) {
|
|
|
|
return Status::InvalidArgument(
|
|
|
|
"unable to parse the specified CF option " + name);
|
|
|
|
}
|
|
|
|
compression_opts.enabled =
|
|
|
|
ParseBoolean("", value.substr(start, value.size() - start));
|
|
|
|
}
|
|
|
|
return Status::OK();
|
|
|
|
}
|
|
|
|
|
Add OptionsUtil::LoadOptionsFromFile() API
Summary:
This patch adds OptionsUtil::LoadOptionsFromFile() and
OptionsUtil::LoadLatestOptionsFromDB(), which allow developers
to construct DBOptions and ColumnFamilyOptions from a RocksDB
options file. Note that most pointer-typed options such as
merge_operator will not be constructed.
With this API, developers no longer need to remember all the
options in order to reopen an existing rocksdb instance like
the following:
DBOptions db_options;
std::vector<std::string> cf_names;
std::vector<ColumnFamilyOptions> cf_opts;
// Load primitive-typed options from an existing DB
OptionsUtil::LoadLatestOptionsFromDB(
dbname, &db_options, &cf_names, &cf_opts);
// Initialize necessary pointer-typed options
cf_opts[0].merge_operator.reset(new MyMergeOperator());
...
// Construct the vector of ColumnFamilyDescriptor
std::vector<ColumnFamilyDescriptor> cf_descs;
for (size_t i = 0; i < cf_opts.size(); ++i) {
cf_descs.emplace_back(cf_names[i], cf_opts[i]);
}
// Open the DB
DB* db = nullptr;
std::vector<ColumnFamilyHandle*> cf_handles;
auto s = DB::Open(db_options, dbname, cf_descs,
&handles, &db);
Test Plan:
Augment existing tests in column_family_test
options_test
db_test
Reviewers: igor, IslamAbdelRahman, sdong, anthony
Reviewed By: anthony
Subscribers: dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D49095
2015-11-12 15:52:43 +01:00
|
|
|
Status ParseColumnFamilyOption(const std::string& name,
|
|
|
|
const std::string& org_value,
|
|
|
|
ColumnFamilyOptions* new_options,
|
|
|
|
bool input_strings_escaped = false) {
|
RocksDB Options file format and its serialization / deserialization.
Summary:
This patch defines the format of RocksDB options file, which
follows the INI file format, and implements functions for its
serialization and deserialization. An example RocksDB options
file can be found in examples/rocksdb_option_file_example.ini.
A typical RocksDB options file has three sections, which are
Version, DBOptions, and more than one CFOptions. The RocksDB
options file in general follows the basic INI file format
with the following extensions / modifications:
* Escaped characters
We escaped the following characters:
- \n -- line feed - new line
- \r -- carriage return
- \\ -- backslash \
- \: -- colon symbol :
- \# -- hash tag #
* Comments
We support # style comments. Comments can appear at the ending
part of a line.
* Statements
A statement is of the form option_name = value.
Each statement contains a '=', where extra white-spaces
are supported. However, we don't support multi-lined statement.
Furthermore, each line can only contain at most one statement.
* Section
Sections are of the form [SecitonTitle "SectionArgument"],
where section argument is optional.
* List
We use colon-separated string to represent a list.
For instance, n1:n2:n3:n4 is a list containing four values.
Below is an example of a RocksDB options file:
[Version]
rocksdb_version=4.0.0
options_file_version=1.0
[DBOptions]
max_open_files=12345
max_background_flushes=301
[CFOptions "default"]
[CFOptions "the second column family"]
[CFOptions "the third column family"]
Test Plan: Added many tests in options_test.cc
Reviewers: igor, IslamAbdelRahman, sdong, anthony
Reviewed By: anthony
Subscribers: maykov, dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D46059
2015-09-29 23:42:40 +02:00
|
|
|
const std::string& value =
|
2015-10-28 18:46:01 +01:00
|
|
|
input_strings_escaped ? UnescapeOptionString(org_value) : org_value;
|
2015-02-20 04:11:14 +01:00
|
|
|
try {
|
2016-09-14 06:12:43 +02:00
|
|
|
if (name == "block_based_table_factory") {
|
2015-02-20 04:11:14 +01:00
|
|
|
// Nested options
|
|
|
|
BlockBasedTableOptions table_opt, base_table_options;
|
2017-07-29 01:23:50 +02:00
|
|
|
BlockBasedTableFactory* block_based_table_factory =
|
|
|
|
static_cast_with_check<BlockBasedTableFactory, TableFactory>(
|
|
|
|
new_options->table_factory.get());
|
2015-02-20 04:11:14 +01:00
|
|
|
if (block_based_table_factory != nullptr) {
|
2015-10-30 23:58:46 +01:00
|
|
|
base_table_options = block_based_table_factory->table_options();
|
2015-02-20 04:11:14 +01:00
|
|
|
}
|
|
|
|
Status table_opt_s = GetBlockBasedTableOptionsFromString(
|
|
|
|
base_table_options, value, &table_opt);
|
|
|
|
if (!table_opt_s.ok()) {
|
Add OptionsUtil::LoadOptionsFromFile() API
Summary:
This patch adds OptionsUtil::LoadOptionsFromFile() and
OptionsUtil::LoadLatestOptionsFromDB(), which allow developers
to construct DBOptions and ColumnFamilyOptions from a RocksDB
options file. Note that most pointer-typed options such as
merge_operator will not be constructed.
With this API, developers no longer need to remember all the
options in order to reopen an existing rocksdb instance like
the following:
DBOptions db_options;
std::vector<std::string> cf_names;
std::vector<ColumnFamilyOptions> cf_opts;
// Load primitive-typed options from an existing DB
OptionsUtil::LoadLatestOptionsFromDB(
dbname, &db_options, &cf_names, &cf_opts);
// Initialize necessary pointer-typed options
cf_opts[0].merge_operator.reset(new MyMergeOperator());
...
// Construct the vector of ColumnFamilyDescriptor
std::vector<ColumnFamilyDescriptor> cf_descs;
for (size_t i = 0; i < cf_opts.size(); ++i) {
cf_descs.emplace_back(cf_names[i], cf_opts[i]);
}
// Open the DB
DB* db = nullptr;
std::vector<ColumnFamilyHandle*> cf_handles;
auto s = DB::Open(db_options, dbname, cf_descs,
&handles, &db);
Test Plan:
Augment existing tests in column_family_test
options_test
db_test
Reviewers: igor, IslamAbdelRahman, sdong, anthony
Reviewed By: anthony
Subscribers: dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D49095
2015-11-12 15:52:43 +01:00
|
|
|
return Status::InvalidArgument(
|
|
|
|
"unable to parse the specified CF option " + name);
|
2015-02-20 04:11:14 +01:00
|
|
|
}
|
|
|
|
new_options->table_factory.reset(NewBlockBasedTableFactory(table_opt));
|
2015-10-28 18:46:01 +01:00
|
|
|
} else if (name == "plain_table_factory") {
|
|
|
|
// Nested options
|
|
|
|
PlainTableOptions table_opt, base_table_options;
|
2017-07-29 01:23:50 +02:00
|
|
|
PlainTableFactory* plain_table_factory =
|
|
|
|
static_cast_with_check<PlainTableFactory, TableFactory>(
|
|
|
|
new_options->table_factory.get());
|
2015-10-28 18:46:01 +01:00
|
|
|
if (plain_table_factory != nullptr) {
|
2015-10-30 23:58:46 +01:00
|
|
|
base_table_options = plain_table_factory->table_options();
|
2015-10-28 18:46:01 +01:00
|
|
|
}
|
|
|
|
Status table_opt_s = GetPlainTableOptionsFromString(
|
|
|
|
base_table_options, value, &table_opt);
|
|
|
|
if (!table_opt_s.ok()) {
|
Add OptionsUtil::LoadOptionsFromFile() API
Summary:
This patch adds OptionsUtil::LoadOptionsFromFile() and
OptionsUtil::LoadLatestOptionsFromDB(), which allow developers
to construct DBOptions and ColumnFamilyOptions from a RocksDB
options file. Note that most pointer-typed options such as
merge_operator will not be constructed.
With this API, developers no longer need to remember all the
options in order to reopen an existing rocksdb instance like
the following:
DBOptions db_options;
std::vector<std::string> cf_names;
std::vector<ColumnFamilyOptions> cf_opts;
// Load primitive-typed options from an existing DB
OptionsUtil::LoadLatestOptionsFromDB(
dbname, &db_options, &cf_names, &cf_opts);
// Initialize necessary pointer-typed options
cf_opts[0].merge_operator.reset(new MyMergeOperator());
...
// Construct the vector of ColumnFamilyDescriptor
std::vector<ColumnFamilyDescriptor> cf_descs;
for (size_t i = 0; i < cf_opts.size(); ++i) {
cf_descs.emplace_back(cf_names[i], cf_opts[i]);
}
// Open the DB
DB* db = nullptr;
std::vector<ColumnFamilyHandle*> cf_handles;
auto s = DB::Open(db_options, dbname, cf_descs,
&handles, &db);
Test Plan:
Augment existing tests in column_family_test
options_test
db_test
Reviewers: igor, IslamAbdelRahman, sdong, anthony
Reviewed By: anthony
Subscribers: dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D49095
2015-11-12 15:52:43 +01:00
|
|
|
return Status::InvalidArgument(
|
|
|
|
"unable to parse the specified CF option " + name);
|
2015-10-28 18:46:01 +01:00
|
|
|
}
|
|
|
|
new_options->table_factory.reset(NewPlainTableFactory(table_opt));
|
2015-11-17 23:29:01 +01:00
|
|
|
} else if (name == "memtable") {
|
|
|
|
std::unique_ptr<MemTableRepFactory> new_mem_factory;
|
|
|
|
Status mem_factory_s =
|
|
|
|
GetMemTableRepFactoryFromString(value, &new_mem_factory);
|
|
|
|
if (!mem_factory_s.ok()) {
|
|
|
|
return Status::InvalidArgument(
|
|
|
|
"unable to parse the specified CF option " + name);
|
|
|
|
}
|
|
|
|
new_options->memtable_factory.reset(new_mem_factory.release());
|
2018-06-28 02:34:07 +02:00
|
|
|
} else if (name == "bottommost_compression_opts") {
|
|
|
|
Status s = ParseCompressionOptions(
|
|
|
|
value, name, new_options->bottommost_compression_opts);
|
|
|
|
if (!s.ok()) {
|
|
|
|
return s;
|
2015-02-20 04:11:14 +01:00
|
|
|
}
|
2018-06-28 02:34:07 +02:00
|
|
|
} else if (name == "compression_opts") {
|
|
|
|
Status s =
|
|
|
|
ParseCompressionOptions(value, name, new_options->compression_opts);
|
|
|
|
if (!s.ok()) {
|
|
|
|
return s;
|
Shared dictionary compression using reference block
Summary:
This adds a new metablock containing a shared dictionary that is used
to compress all data blocks in the SST file. The size of the shared dictionary
is configurable in CompressionOptions and defaults to 0. It's currently only
used for zlib/lz4/lz4hc, but the block will be stored in the SST regardless of
the compression type if the user chooses a nonzero dictionary size.
During compaction, computes the dictionary by randomly sampling the first
output file in each subcompaction. It pre-computes the intervals to sample
by assuming the output file will have the maximum allowable length. In case
the file is smaller, some of the pre-computed sampling intervals can be beyond
end-of-file, in which case we skip over those samples and the dictionary will
be a bit smaller. After the dictionary is generated using the first file in a
subcompaction, it is loaded into the compression library before writing each
block in each subsequent file of that subcompaction.
On the read path, gets the dictionary from the metablock, if it exists. Then,
loads that dictionary into the compression library before reading each block.
Test Plan: new unit test
Reviewers: yhchiang, IslamAbdelRahman, cyan, sdong
Reviewed By: sdong
Subscribers: andrewkr, yoshinorim, kradhakrishnan, dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D52287
2016-04-28 02:36:03 +02:00
|
|
|
}
|
2015-02-20 04:11:14 +01:00
|
|
|
} else {
|
2019-03-26 22:15:26 +01:00
|
|
|
if (name == kNameComparator) {
|
|
|
|
// Try to get comparator from object registry first.
|
|
|
|
// Only support static comparator for now.
|
2019-07-24 02:08:26 +02:00
|
|
|
Status status = ObjectRegistry::NewInstance()->NewStaticObject(
|
|
|
|
value, &new_options->comparator);
|
|
|
|
if (status.ok()) {
|
|
|
|
return status;
|
2019-03-26 22:15:26 +01:00
|
|
|
}
|
2019-03-28 22:50:06 +01:00
|
|
|
} else if (name == kNameMergeOperator) {
|
|
|
|
// Try to get merge operator from object registry first.
|
2019-07-24 02:08:26 +02:00
|
|
|
std::shared_ptr<MergeOperator> mo;
|
|
|
|
Status status =
|
|
|
|
ObjectRegistry::NewInstance()->NewSharedObject<MergeOperator>(
|
|
|
|
value, &new_options->merge_operator);
|
2019-03-28 22:50:06 +01:00
|
|
|
// Only support static comparator for now.
|
2019-07-24 02:08:26 +02:00
|
|
|
if (status.ok()) {
|
|
|
|
return status;
|
2019-03-28 22:50:06 +01:00
|
|
|
}
|
2019-03-26 22:15:26 +01:00
|
|
|
}
|
2019-03-28 22:50:06 +01:00
|
|
|
|
2015-08-27 01:13:56 +02:00
|
|
|
auto iter = cf_options_type_info.find(name);
|
|
|
|
if (iter == cf_options_type_info.end()) {
|
Add OptionsUtil::LoadOptionsFromFile() API
Summary:
This patch adds OptionsUtil::LoadOptionsFromFile() and
OptionsUtil::LoadLatestOptionsFromDB(), which allow developers
to construct DBOptions and ColumnFamilyOptions from a RocksDB
options file. Note that most pointer-typed options such as
merge_operator will not be constructed.
With this API, developers no longer need to remember all the
options in order to reopen an existing rocksdb instance like
the following:
DBOptions db_options;
std::vector<std::string> cf_names;
std::vector<ColumnFamilyOptions> cf_opts;
// Load primitive-typed options from an existing DB
OptionsUtil::LoadLatestOptionsFromDB(
dbname, &db_options, &cf_names, &cf_opts);
// Initialize necessary pointer-typed options
cf_opts[0].merge_operator.reset(new MyMergeOperator());
...
// Construct the vector of ColumnFamilyDescriptor
std::vector<ColumnFamilyDescriptor> cf_descs;
for (size_t i = 0; i < cf_opts.size(); ++i) {
cf_descs.emplace_back(cf_names[i], cf_opts[i]);
}
// Open the DB
DB* db = nullptr;
std::vector<ColumnFamilyHandle*> cf_handles;
auto s = DB::Open(db_options, dbname, cf_descs,
&handles, &db);
Test Plan:
Augment existing tests in column_family_test
options_test
db_test
Reviewers: igor, IslamAbdelRahman, sdong, anthony
Reviewed By: anthony
Subscribers: dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D49095
2015-11-12 15:52:43 +01:00
|
|
|
return Status::InvalidArgument(
|
|
|
|
"Unable to parse the specified CF option " + name);
|
2015-08-27 01:13:56 +02:00
|
|
|
}
|
|
|
|
const auto& opt_info = iter->second;
|
2016-06-17 13:52:53 +02:00
|
|
|
if (opt_info.verification != OptionVerificationType::kDeprecated &&
|
|
|
|
ParseOptionHelper(
|
Add OptionsUtil::LoadOptionsFromFile() API
Summary:
This patch adds OptionsUtil::LoadOptionsFromFile() and
OptionsUtil::LoadLatestOptionsFromDB(), which allow developers
to construct DBOptions and ColumnFamilyOptions from a RocksDB
options file. Note that most pointer-typed options such as
merge_operator will not be constructed.
With this API, developers no longer need to remember all the
options in order to reopen an existing rocksdb instance like
the following:
DBOptions db_options;
std::vector<std::string> cf_names;
std::vector<ColumnFamilyOptions> cf_opts;
// Load primitive-typed options from an existing DB
OptionsUtil::LoadLatestOptionsFromDB(
dbname, &db_options, &cf_names, &cf_opts);
// Initialize necessary pointer-typed options
cf_opts[0].merge_operator.reset(new MyMergeOperator());
...
// Construct the vector of ColumnFamilyDescriptor
std::vector<ColumnFamilyDescriptor> cf_descs;
for (size_t i = 0; i < cf_opts.size(); ++i) {
cf_descs.emplace_back(cf_names[i], cf_opts[i]);
}
// Open the DB
DB* db = nullptr;
std::vector<ColumnFamilyHandle*> cf_handles;
auto s = DB::Open(db_options, dbname, cf_descs,
&handles, &db);
Test Plan:
Augment existing tests in column_family_test
options_test
db_test
Reviewers: igor, IslamAbdelRahman, sdong, anthony
Reviewed By: anthony
Subscribers: dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D49095
2015-11-12 15:52:43 +01:00
|
|
|
reinterpret_cast<char*>(new_options) + opt_info.offset,
|
|
|
|
opt_info.type, value)) {
|
|
|
|
return Status::OK();
|
|
|
|
}
|
|
|
|
switch (opt_info.verification) {
|
|
|
|
case OptionVerificationType::kByName:
|
2016-02-19 23:42:24 +01:00
|
|
|
case OptionVerificationType::kByNameAllowNull:
|
2017-10-04 18:44:18 +02:00
|
|
|
case OptionVerificationType::kByNameAllowFromNull:
|
Add OptionsUtil::LoadOptionsFromFile() API
Summary:
This patch adds OptionsUtil::LoadOptionsFromFile() and
OptionsUtil::LoadLatestOptionsFromDB(), which allow developers
to construct DBOptions and ColumnFamilyOptions from a RocksDB
options file. Note that most pointer-typed options such as
merge_operator will not be constructed.
With this API, developers no longer need to remember all the
options in order to reopen an existing rocksdb instance like
the following:
DBOptions db_options;
std::vector<std::string> cf_names;
std::vector<ColumnFamilyOptions> cf_opts;
// Load primitive-typed options from an existing DB
OptionsUtil::LoadLatestOptionsFromDB(
dbname, &db_options, &cf_names, &cf_opts);
// Initialize necessary pointer-typed options
cf_opts[0].merge_operator.reset(new MyMergeOperator());
...
// Construct the vector of ColumnFamilyDescriptor
std::vector<ColumnFamilyDescriptor> cf_descs;
for (size_t i = 0; i < cf_opts.size(); ++i) {
cf_descs.emplace_back(cf_names[i], cf_opts[i]);
}
// Open the DB
DB* db = nullptr;
std::vector<ColumnFamilyHandle*> cf_handles;
auto s = DB::Open(db_options, dbname, cf_descs,
&handles, &db);
Test Plan:
Augment existing tests in column_family_test
options_test
db_test
Reviewers: igor, IslamAbdelRahman, sdong, anthony
Reviewed By: anthony
Subscribers: dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D49095
2015-11-12 15:52:43 +01:00
|
|
|
return Status::NotSupported(
|
|
|
|
"Deserializing the specified CF option " + name +
|
|
|
|
" is not supported");
|
|
|
|
case OptionVerificationType::kDeprecated:
|
|
|
|
return Status::OK();
|
|
|
|
default:
|
|
|
|
return Status::InvalidArgument(
|
|
|
|
"Unable to parse the specified CF option " + name);
|
|
|
|
}
|
2015-02-20 04:11:14 +01:00
|
|
|
}
|
2015-11-02 23:11:28 +01:00
|
|
|
} catch (const std::exception&) {
|
Add OptionsUtil::LoadOptionsFromFile() API
Summary:
This patch adds OptionsUtil::LoadOptionsFromFile() and
OptionsUtil::LoadLatestOptionsFromDB(), which allow developers
to construct DBOptions and ColumnFamilyOptions from a RocksDB
options file. Note that most pointer-typed options such as
merge_operator will not be constructed.
With this API, developers no longer need to remember all the
options in order to reopen an existing rocksdb instance like
the following:
DBOptions db_options;
std::vector<std::string> cf_names;
std::vector<ColumnFamilyOptions> cf_opts;
// Load primitive-typed options from an existing DB
OptionsUtil::LoadLatestOptionsFromDB(
dbname, &db_options, &cf_names, &cf_opts);
// Initialize necessary pointer-typed options
cf_opts[0].merge_operator.reset(new MyMergeOperator());
...
// Construct the vector of ColumnFamilyDescriptor
std::vector<ColumnFamilyDescriptor> cf_descs;
for (size_t i = 0; i < cf_opts.size(); ++i) {
cf_descs.emplace_back(cf_names[i], cf_opts[i]);
}
// Open the DB
DB* db = nullptr;
std::vector<ColumnFamilyHandle*> cf_handles;
auto s = DB::Open(db_options, dbname, cf_descs,
&handles, &db);
Test Plan:
Augment existing tests in column_family_test
options_test
db_test
Reviewers: igor, IslamAbdelRahman, sdong, anthony
Reviewed By: anthony
Subscribers: dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D49095
2015-11-12 15:52:43 +01:00
|
|
|
return Status::InvalidArgument(
|
|
|
|
"unable to parse the specified option " + name);
|
2015-02-20 04:11:14 +01:00
|
|
|
}
|
Add OptionsUtil::LoadOptionsFromFile() API
Summary:
This patch adds OptionsUtil::LoadOptionsFromFile() and
OptionsUtil::LoadLatestOptionsFromDB(), which allow developers
to construct DBOptions and ColumnFamilyOptions from a RocksDB
options file. Note that most pointer-typed options such as
merge_operator will not be constructed.
With this API, developers no longer need to remember all the
options in order to reopen an existing rocksdb instance like
the following:
DBOptions db_options;
std::vector<std::string> cf_names;
std::vector<ColumnFamilyOptions> cf_opts;
// Load primitive-typed options from an existing DB
OptionsUtil::LoadLatestOptionsFromDB(
dbname, &db_options, &cf_names, &cf_opts);
// Initialize necessary pointer-typed options
cf_opts[0].merge_operator.reset(new MyMergeOperator());
...
// Construct the vector of ColumnFamilyDescriptor
std::vector<ColumnFamilyDescriptor> cf_descs;
for (size_t i = 0; i < cf_opts.size(); ++i) {
cf_descs.emplace_back(cf_names[i], cf_opts[i]);
}
// Open the DB
DB* db = nullptr;
std::vector<ColumnFamilyHandle*> cf_handles;
auto s = DB::Open(db_options, dbname, cf_descs,
&handles, &db);
Test Plan:
Augment existing tests in column_family_test
options_test
db_test
Reviewers: igor, IslamAbdelRahman, sdong, anthony
Reviewed By: anthony
Subscribers: dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D49095
2015-11-12 15:52:43 +01:00
|
|
|
return Status::OK();
|
2015-02-20 04:11:14 +01:00
|
|
|
}
|
|
|
|
|
2017-10-20 00:19:20 +02:00
|
|
|
template <typename T>
|
|
|
|
bool SerializeSingleStructOption(
|
|
|
|
std::string* opt_string, const T& options,
|
2019-05-15 23:19:04 +02:00
|
|
|
const std::unordered_map<std::string, OptionTypeInfo>& type_info,
|
2017-10-20 00:19:20 +02:00
|
|
|
const std::string& name, const std::string& delimiter) {
|
|
|
|
auto iter = type_info.find(name);
|
|
|
|
if (iter == type_info.end()) {
|
2015-08-18 22:30:18 +02:00
|
|
|
return false;
|
|
|
|
}
|
|
|
|
auto& opt_info = iter->second;
|
|
|
|
const char* opt_address =
|
2017-10-20 00:19:20 +02:00
|
|
|
reinterpret_cast<const char*>(&options) + opt_info.offset;
|
2015-08-18 22:30:18 +02:00
|
|
|
std::string value;
|
2015-08-27 01:13:56 +02:00
|
|
|
bool result = SerializeSingleOptionHelper(opt_address, opt_info.type, &value);
|
|
|
|
if (result) {
|
RocksDB Options file format and its serialization / deserialization.
Summary:
This patch defines the format of RocksDB options file, which
follows the INI file format, and implements functions for its
serialization and deserialization. An example RocksDB options
file can be found in examples/rocksdb_option_file_example.ini.
A typical RocksDB options file has three sections, which are
Version, DBOptions, and more than one CFOptions. The RocksDB
options file in general follows the basic INI file format
with the following extensions / modifications:
* Escaped characters
We escaped the following characters:
- \n -- line feed - new line
- \r -- carriage return
- \\ -- backslash \
- \: -- colon symbol :
- \# -- hash tag #
* Comments
We support # style comments. Comments can appear at the ending
part of a line.
* Statements
A statement is of the form option_name = value.
Each statement contains a '=', where extra white-spaces
are supported. However, we don't support multi-lined statement.
Furthermore, each line can only contain at most one statement.
* Section
Sections are of the form [SecitonTitle "SectionArgument"],
where section argument is optional.
* List
We use colon-separated string to represent a list.
For instance, n1:n2:n3:n4 is a list containing four values.
Below is an example of a RocksDB options file:
[Version]
rocksdb_version=4.0.0
options_file_version=1.0
[DBOptions]
max_open_files=12345
max_background_flushes=301
[CFOptions "default"]
[CFOptions "the second column family"]
[CFOptions "the third column family"]
Test Plan: Added many tests in options_test.cc
Reviewers: igor, IslamAbdelRahman, sdong, anthony
Reviewed By: anthony
Subscribers: maykov, dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D46059
2015-09-29 23:42:40 +02:00
|
|
|
*opt_string = name + "=" + value + delimiter;
|
2015-08-18 22:30:18 +02:00
|
|
|
}
|
2015-08-27 01:13:56 +02:00
|
|
|
return result;
|
2015-08-18 22:30:18 +02:00
|
|
|
}
|
|
|
|
|
2017-10-20 00:19:20 +02:00
|
|
|
template <typename T>
|
|
|
|
Status GetStringFromStruct(
|
|
|
|
std::string* opt_string, const T& options,
|
2019-05-15 23:19:04 +02:00
|
|
|
const std::unordered_map<std::string, OptionTypeInfo>& type_info,
|
2017-10-20 00:19:20 +02:00
|
|
|
const std::string& delimiter) {
|
2015-08-18 22:30:18 +02:00
|
|
|
assert(opt_string);
|
|
|
|
opt_string->clear();
|
2017-10-20 00:19:20 +02:00
|
|
|
for (auto iter = type_info.begin(); iter != type_info.end(); ++iter) {
|
2015-09-30 02:58:00 +02:00
|
|
|
if (iter->second.verification == OptionVerificationType::kDeprecated) {
|
|
|
|
// If the option is no longer used in rocksdb and marked as deprecated,
|
|
|
|
// we skip it in the serialization.
|
|
|
|
continue;
|
|
|
|
}
|
2015-08-18 22:30:18 +02:00
|
|
|
std::string single_output;
|
2017-10-20 00:19:20 +02:00
|
|
|
bool result = SerializeSingleStructOption<T>(
|
|
|
|
&single_output, options, type_info, iter->first, delimiter);
|
2015-08-18 22:30:18 +02:00
|
|
|
if (result) {
|
|
|
|
opt_string->append(single_output);
|
2017-10-20 00:19:20 +02:00
|
|
|
} else {
|
|
|
|
return Status::InvalidArgument("failed to serialize %s\n",
|
|
|
|
iter->first.c_str());
|
2015-08-18 22:30:18 +02:00
|
|
|
}
|
2017-10-20 00:19:20 +02:00
|
|
|
assert(result);
|
2015-08-18 22:30:18 +02:00
|
|
|
}
|
|
|
|
return Status::OK();
|
|
|
|
}
|
|
|
|
|
2017-10-20 00:19:20 +02:00
|
|
|
Status GetStringFromDBOptions(std::string* opt_string,
|
|
|
|
const DBOptions& db_options,
|
|
|
|
const std::string& delimiter) {
|
|
|
|
return GetStringFromStruct<DBOptions>(opt_string, db_options,
|
|
|
|
db_options_type_info, delimiter);
|
2015-08-27 01:13:56 +02:00
|
|
|
}
|
|
|
|
|
RocksDB Options file format and its serialization / deserialization.
Summary:
This patch defines the format of RocksDB options file, which
follows the INI file format, and implements functions for its
serialization and deserialization. An example RocksDB options
file can be found in examples/rocksdb_option_file_example.ini.
A typical RocksDB options file has three sections, which are
Version, DBOptions, and more than one CFOptions. The RocksDB
options file in general follows the basic INI file format
with the following extensions / modifications:
* Escaped characters
We escaped the following characters:
- \n -- line feed - new line
- \r -- carriage return
- \\ -- backslash \
- \: -- colon symbol :
- \# -- hash tag #
* Comments
We support # style comments. Comments can appear at the ending
part of a line.
* Statements
A statement is of the form option_name = value.
Each statement contains a '=', where extra white-spaces
are supported. However, we don't support multi-lined statement.
Furthermore, each line can only contain at most one statement.
* Section
Sections are of the form [SecitonTitle "SectionArgument"],
where section argument is optional.
* List
We use colon-separated string to represent a list.
For instance, n1:n2:n3:n4 is a list containing four values.
Below is an example of a RocksDB options file:
[Version]
rocksdb_version=4.0.0
options_file_version=1.0
[DBOptions]
max_open_files=12345
max_background_flushes=301
[CFOptions "default"]
[CFOptions "the second column family"]
[CFOptions "the third column family"]
Test Plan: Added many tests in options_test.cc
Reviewers: igor, IslamAbdelRahman, sdong, anthony
Reviewed By: anthony
Subscribers: maykov, dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D46059
2015-09-29 23:42:40 +02:00
|
|
|
Status GetStringFromColumnFamilyOptions(std::string* opt_string,
|
|
|
|
const ColumnFamilyOptions& cf_options,
|
|
|
|
const std::string& delimiter) {
|
2017-10-20 00:19:20 +02:00
|
|
|
return GetStringFromStruct<ColumnFamilyOptions>(
|
|
|
|
opt_string, cf_options, cf_options_type_info, delimiter);
|
2015-08-27 01:13:56 +02:00
|
|
|
}
|
|
|
|
|
2016-05-19 00:03:21 +02:00
|
|
|
Status GetStringFromCompressionType(std::string* compression_str,
|
|
|
|
CompressionType compression_type) {
|
|
|
|
bool ok = SerializeEnum<CompressionType>(compression_type_string_map,
|
|
|
|
compression_type, compression_str);
|
|
|
|
if (ok) {
|
|
|
|
return Status::OK();
|
|
|
|
} else {
|
|
|
|
return Status::InvalidArgument("Invalid compression types");
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2017-01-06 20:11:51 +01:00
|
|
|
std::vector<CompressionType> GetSupportedCompressions() {
|
|
|
|
std::vector<CompressionType> supported_compressions;
|
|
|
|
for (const auto& comp_to_name : compression_type_string_map) {
|
|
|
|
CompressionType t = comp_to_name.second;
|
|
|
|
if (t != kDisableCompressionOption && CompressionTypeSupported(t)) {
|
|
|
|
supported_compressions.push_back(t);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
return supported_compressions;
|
|
|
|
}
|
|
|
|
|
Add OptionsUtil::LoadOptionsFromFile() API
Summary:
This patch adds OptionsUtil::LoadOptionsFromFile() and
OptionsUtil::LoadLatestOptionsFromDB(), which allow developers
to construct DBOptions and ColumnFamilyOptions from a RocksDB
options file. Note that most pointer-typed options such as
merge_operator will not be constructed.
With this API, developers no longer need to remember all the
options in order to reopen an existing rocksdb instance like
the following:
DBOptions db_options;
std::vector<std::string> cf_names;
std::vector<ColumnFamilyOptions> cf_opts;
// Load primitive-typed options from an existing DB
OptionsUtil::LoadLatestOptionsFromDB(
dbname, &db_options, &cf_names, &cf_opts);
// Initialize necessary pointer-typed options
cf_opts[0].merge_operator.reset(new MyMergeOperator());
...
// Construct the vector of ColumnFamilyDescriptor
std::vector<ColumnFamilyDescriptor> cf_descs;
for (size_t i = 0; i < cf_opts.size(); ++i) {
cf_descs.emplace_back(cf_names[i], cf_opts[i]);
}
// Open the DB
DB* db = nullptr;
std::vector<ColumnFamilyHandle*> cf_handles;
auto s = DB::Open(db_options, dbname, cf_descs,
&handles, &db);
Test Plan:
Augment existing tests in column_family_test
options_test
db_test
Reviewers: igor, IslamAbdelRahman, sdong, anthony
Reviewed By: anthony
Subscribers: dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D49095
2015-11-12 15:52:43 +01:00
|
|
|
Status ParseDBOption(const std::string& name,
|
|
|
|
const std::string& org_value,
|
|
|
|
DBOptions* new_options,
|
|
|
|
bool input_strings_escaped = false) {
|
RocksDB Options file format and its serialization / deserialization.
Summary:
This patch defines the format of RocksDB options file, which
follows the INI file format, and implements functions for its
serialization and deserialization. An example RocksDB options
file can be found in examples/rocksdb_option_file_example.ini.
A typical RocksDB options file has three sections, which are
Version, DBOptions, and more than one CFOptions. The RocksDB
options file in general follows the basic INI file format
with the following extensions / modifications:
* Escaped characters
We escaped the following characters:
- \n -- line feed - new line
- \r -- carriage return
- \\ -- backslash \
- \: -- colon symbol :
- \# -- hash tag #
* Comments
We support # style comments. Comments can appear at the ending
part of a line.
* Statements
A statement is of the form option_name = value.
Each statement contains a '=', where extra white-spaces
are supported. However, we don't support multi-lined statement.
Furthermore, each line can only contain at most one statement.
* Section
Sections are of the form [SecitonTitle "SectionArgument"],
where section argument is optional.
* List
We use colon-separated string to represent a list.
For instance, n1:n2:n3:n4 is a list containing four values.
Below is an example of a RocksDB options file:
[Version]
rocksdb_version=4.0.0
options_file_version=1.0
[DBOptions]
max_open_files=12345
max_background_flushes=301
[CFOptions "default"]
[CFOptions "the second column family"]
[CFOptions "the third column family"]
Test Plan: Added many tests in options_test.cc
Reviewers: igor, IslamAbdelRahman, sdong, anthony
Reviewed By: anthony
Subscribers: maykov, dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D46059
2015-09-29 23:42:40 +02:00
|
|
|
const std::string& value =
|
2015-10-28 18:46:01 +01:00
|
|
|
input_strings_escaped ? UnescapeOptionString(org_value) : org_value;
|
2015-02-20 04:11:14 +01:00
|
|
|
try {
|
2015-08-18 22:30:18 +02:00
|
|
|
if (name == "rate_limiter_bytes_per_sec") {
|
2015-03-06 23:21:15 +01:00
|
|
|
new_options->rate_limiter.reset(
|
|
|
|
NewGenericRateLimiter(static_cast<int64_t>(ParseUint64(value))));
|
2019-04-25 20:31:58 +02:00
|
|
|
} else if (name == kNameEnv) {
|
|
|
|
// Currently `Env` can be deserialized from object registry only.
|
2019-07-24 02:08:26 +02:00
|
|
|
Env* env = new_options->env;
|
|
|
|
Status status = Env::LoadEnv(value, &env);
|
2019-04-25 20:31:58 +02:00
|
|
|
// Only support static env for now.
|
2019-07-24 02:08:26 +02:00
|
|
|
if (status.ok()) {
|
2019-04-25 20:31:58 +02:00
|
|
|
new_options->env = env;
|
|
|
|
}
|
2015-02-20 04:11:14 +01:00
|
|
|
} else {
|
2015-08-18 22:30:18 +02:00
|
|
|
auto iter = db_options_type_info.find(name);
|
|
|
|
if (iter == db_options_type_info.end()) {
|
Add OptionsUtil::LoadOptionsFromFile() API
Summary:
This patch adds OptionsUtil::LoadOptionsFromFile() and
OptionsUtil::LoadLatestOptionsFromDB(), which allow developers
to construct DBOptions and ColumnFamilyOptions from a RocksDB
options file. Note that most pointer-typed options such as
merge_operator will not be constructed.
With this API, developers no longer need to remember all the
options in order to reopen an existing rocksdb instance like
the following:
DBOptions db_options;
std::vector<std::string> cf_names;
std::vector<ColumnFamilyOptions> cf_opts;
// Load primitive-typed options from an existing DB
OptionsUtil::LoadLatestOptionsFromDB(
dbname, &db_options, &cf_names, &cf_opts);
// Initialize necessary pointer-typed options
cf_opts[0].merge_operator.reset(new MyMergeOperator());
...
// Construct the vector of ColumnFamilyDescriptor
std::vector<ColumnFamilyDescriptor> cf_descs;
for (size_t i = 0; i < cf_opts.size(); ++i) {
cf_descs.emplace_back(cf_names[i], cf_opts[i]);
}
// Open the DB
DB* db = nullptr;
std::vector<ColumnFamilyHandle*> cf_handles;
auto s = DB::Open(db_options, dbname, cf_descs,
&handles, &db);
Test Plan:
Augment existing tests in column_family_test
options_test
db_test
Reviewers: igor, IslamAbdelRahman, sdong, anthony
Reviewed By: anthony
Subscribers: dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D49095
2015-11-12 15:52:43 +01:00
|
|
|
return Status::InvalidArgument("Unrecognized option DBOptions:", name);
|
2015-08-18 22:30:18 +02:00
|
|
|
}
|
2015-08-27 01:13:56 +02:00
|
|
|
const auto& opt_info = iter->second;
|
2016-06-17 13:52:53 +02:00
|
|
|
if (opt_info.verification != OptionVerificationType::kDeprecated &&
|
|
|
|
ParseOptionHelper(
|
Add OptionsUtil::LoadOptionsFromFile() API
Summary:
This patch adds OptionsUtil::LoadOptionsFromFile() and
OptionsUtil::LoadLatestOptionsFromDB(), which allow developers
to construct DBOptions and ColumnFamilyOptions from a RocksDB
options file. Note that most pointer-typed options such as
merge_operator will not be constructed.
With this API, developers no longer need to remember all the
options in order to reopen an existing rocksdb instance like
the following:
DBOptions db_options;
std::vector<std::string> cf_names;
std::vector<ColumnFamilyOptions> cf_opts;
// Load primitive-typed options from an existing DB
OptionsUtil::LoadLatestOptionsFromDB(
dbname, &db_options, &cf_names, &cf_opts);
// Initialize necessary pointer-typed options
cf_opts[0].merge_operator.reset(new MyMergeOperator());
...
// Construct the vector of ColumnFamilyDescriptor
std::vector<ColumnFamilyDescriptor> cf_descs;
for (size_t i = 0; i < cf_opts.size(); ++i) {
cf_descs.emplace_back(cf_names[i], cf_opts[i]);
}
// Open the DB
DB* db = nullptr;
std::vector<ColumnFamilyHandle*> cf_handles;
auto s = DB::Open(db_options, dbname, cf_descs,
&handles, &db);
Test Plan:
Augment existing tests in column_family_test
options_test
db_test
Reviewers: igor, IslamAbdelRahman, sdong, anthony
Reviewed By: anthony
Subscribers: dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D49095
2015-11-12 15:52:43 +01:00
|
|
|
reinterpret_cast<char*>(new_options) + opt_info.offset,
|
|
|
|
opt_info.type, value)) {
|
|
|
|
return Status::OK();
|
|
|
|
}
|
|
|
|
switch (opt_info.verification) {
|
|
|
|
case OptionVerificationType::kByName:
|
2016-02-19 23:42:24 +01:00
|
|
|
case OptionVerificationType::kByNameAllowNull:
|
Add OptionsUtil::LoadOptionsFromFile() API
Summary:
This patch adds OptionsUtil::LoadOptionsFromFile() and
OptionsUtil::LoadLatestOptionsFromDB(), which allow developers
to construct DBOptions and ColumnFamilyOptions from a RocksDB
options file. Note that most pointer-typed options such as
merge_operator will not be constructed.
With this API, developers no longer need to remember all the
options in order to reopen an existing rocksdb instance like
the following:
DBOptions db_options;
std::vector<std::string> cf_names;
std::vector<ColumnFamilyOptions> cf_opts;
// Load primitive-typed options from an existing DB
OptionsUtil::LoadLatestOptionsFromDB(
dbname, &db_options, &cf_names, &cf_opts);
// Initialize necessary pointer-typed options
cf_opts[0].merge_operator.reset(new MyMergeOperator());
...
// Construct the vector of ColumnFamilyDescriptor
std::vector<ColumnFamilyDescriptor> cf_descs;
for (size_t i = 0; i < cf_opts.size(); ++i) {
cf_descs.emplace_back(cf_names[i], cf_opts[i]);
}
// Open the DB
DB* db = nullptr;
std::vector<ColumnFamilyHandle*> cf_handles;
auto s = DB::Open(db_options, dbname, cf_descs,
&handles, &db);
Test Plan:
Augment existing tests in column_family_test
options_test
db_test
Reviewers: igor, IslamAbdelRahman, sdong, anthony
Reviewed By: anthony
Subscribers: dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D49095
2015-11-12 15:52:43 +01:00
|
|
|
return Status::NotSupported(
|
|
|
|
"Deserializing the specified DB option " + name +
|
|
|
|
" is not supported");
|
|
|
|
case OptionVerificationType::kDeprecated:
|
|
|
|
return Status::OK();
|
|
|
|
default:
|
|
|
|
return Status::InvalidArgument(
|
|
|
|
"Unable to parse the specified DB option " + name);
|
2015-10-03 00:35:32 +02:00
|
|
|
}
|
2015-02-20 04:11:14 +01:00
|
|
|
}
|
2015-11-02 23:11:28 +01:00
|
|
|
} catch (const std::exception&) {
|
Add OptionsUtil::LoadOptionsFromFile() API
Summary:
This patch adds OptionsUtil::LoadOptionsFromFile() and
OptionsUtil::LoadLatestOptionsFromDB(), which allow developers
to construct DBOptions and ColumnFamilyOptions from a RocksDB
options file. Note that most pointer-typed options such as
merge_operator will not be constructed.
With this API, developers no longer need to remember all the
options in order to reopen an existing rocksdb instance like
the following:
DBOptions db_options;
std::vector<std::string> cf_names;
std::vector<ColumnFamilyOptions> cf_opts;
// Load primitive-typed options from an existing DB
OptionsUtil::LoadLatestOptionsFromDB(
dbname, &db_options, &cf_names, &cf_opts);
// Initialize necessary pointer-typed options
cf_opts[0].merge_operator.reset(new MyMergeOperator());
...
// Construct the vector of ColumnFamilyDescriptor
std::vector<ColumnFamilyDescriptor> cf_descs;
for (size_t i = 0; i < cf_opts.size(); ++i) {
cf_descs.emplace_back(cf_names[i], cf_opts[i]);
}
// Open the DB
DB* db = nullptr;
std::vector<ColumnFamilyHandle*> cf_handles;
auto s = DB::Open(db_options, dbname, cf_descs,
&handles, &db);
Test Plan:
Augment existing tests in column_family_test
options_test
db_test
Reviewers: igor, IslamAbdelRahman, sdong, anthony
Reviewed By: anthony
Subscribers: dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D49095
2015-11-12 15:52:43 +01:00
|
|
|
return Status::InvalidArgument("Unable to parse DBOptions:", name);
|
2015-02-20 04:11:14 +01:00
|
|
|
}
|
Add OptionsUtil::LoadOptionsFromFile() API
Summary:
This patch adds OptionsUtil::LoadOptionsFromFile() and
OptionsUtil::LoadLatestOptionsFromDB(), which allow developers
to construct DBOptions and ColumnFamilyOptions from a RocksDB
options file. Note that most pointer-typed options such as
merge_operator will not be constructed.
With this API, developers no longer need to remember all the
options in order to reopen an existing rocksdb instance like
the following:
DBOptions db_options;
std::vector<std::string> cf_names;
std::vector<ColumnFamilyOptions> cf_opts;
// Load primitive-typed options from an existing DB
OptionsUtil::LoadLatestOptionsFromDB(
dbname, &db_options, &cf_names, &cf_opts);
// Initialize necessary pointer-typed options
cf_opts[0].merge_operator.reset(new MyMergeOperator());
...
// Construct the vector of ColumnFamilyDescriptor
std::vector<ColumnFamilyDescriptor> cf_descs;
for (size_t i = 0; i < cf_opts.size(); ++i) {
cf_descs.emplace_back(cf_names[i], cf_opts[i]);
}
// Open the DB
DB* db = nullptr;
std::vector<ColumnFamilyHandle*> cf_handles;
auto s = DB::Open(db_options, dbname, cf_descs,
&handles, &db);
Test Plan:
Augment existing tests in column_family_test
options_test
db_test
Reviewers: igor, IslamAbdelRahman, sdong, anthony
Reviewed By: anthony
Subscribers: dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D49095
2015-11-12 15:52:43 +01:00
|
|
|
return Status::OK();
|
2015-02-20 04:11:14 +01:00
|
|
|
}
|
2014-10-10 19:00:12 +02:00
|
|
|
|
2014-12-22 22:18:57 +01:00
|
|
|
Status GetColumnFamilyOptionsFromMap(
|
2014-10-10 19:00:12 +02:00
|
|
|
const ColumnFamilyOptions& base_options,
|
|
|
|
const std::unordered_map<std::string, std::string>& opts_map,
|
2017-06-14 01:55:08 +02:00
|
|
|
ColumnFamilyOptions* new_options, bool input_strings_escaped,
|
|
|
|
bool ignore_unknown_options) {
|
Add OptionsUtil::LoadOptionsFromFile() API
Summary:
This patch adds OptionsUtil::LoadOptionsFromFile() and
OptionsUtil::LoadLatestOptionsFromDB(), which allow developers
to construct DBOptions and ColumnFamilyOptions from a RocksDB
options file. Note that most pointer-typed options such as
merge_operator will not be constructed.
With this API, developers no longer need to remember all the
options in order to reopen an existing rocksdb instance like
the following:
DBOptions db_options;
std::vector<std::string> cf_names;
std::vector<ColumnFamilyOptions> cf_opts;
// Load primitive-typed options from an existing DB
OptionsUtil::LoadLatestOptionsFromDB(
dbname, &db_options, &cf_names, &cf_opts);
// Initialize necessary pointer-typed options
cf_opts[0].merge_operator.reset(new MyMergeOperator());
...
// Construct the vector of ColumnFamilyDescriptor
std::vector<ColumnFamilyDescriptor> cf_descs;
for (size_t i = 0; i < cf_opts.size(); ++i) {
cf_descs.emplace_back(cf_names[i], cf_opts[i]);
}
// Open the DB
DB* db = nullptr;
std::vector<ColumnFamilyHandle*> cf_handles;
auto s = DB::Open(db_options, dbname, cf_descs,
&handles, &db);
Test Plan:
Augment existing tests in column_family_test
options_test
db_test
Reviewers: igor, IslamAbdelRahman, sdong, anthony
Reviewed By: anthony
Subscribers: dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D49095
2015-11-12 15:52:43 +01:00
|
|
|
return GetColumnFamilyOptionsFromMapInternal(
|
2017-06-14 01:55:08 +02:00
|
|
|
base_options, opts_map, new_options, input_strings_escaped, nullptr,
|
|
|
|
ignore_unknown_options);
|
Add OptionsUtil::LoadOptionsFromFile() API
Summary:
This patch adds OptionsUtil::LoadOptionsFromFile() and
OptionsUtil::LoadLatestOptionsFromDB(), which allow developers
to construct DBOptions and ColumnFamilyOptions from a RocksDB
options file. Note that most pointer-typed options such as
merge_operator will not be constructed.
With this API, developers no longer need to remember all the
options in order to reopen an existing rocksdb instance like
the following:
DBOptions db_options;
std::vector<std::string> cf_names;
std::vector<ColumnFamilyOptions> cf_opts;
// Load primitive-typed options from an existing DB
OptionsUtil::LoadLatestOptionsFromDB(
dbname, &db_options, &cf_names, &cf_opts);
// Initialize necessary pointer-typed options
cf_opts[0].merge_operator.reset(new MyMergeOperator());
...
// Construct the vector of ColumnFamilyDescriptor
std::vector<ColumnFamilyDescriptor> cf_descs;
for (size_t i = 0; i < cf_opts.size(); ++i) {
cf_descs.emplace_back(cf_names[i], cf_opts[i]);
}
// Open the DB
DB* db = nullptr;
std::vector<ColumnFamilyHandle*> cf_handles;
auto s = DB::Open(db_options, dbname, cf_descs,
&handles, &db);
Test Plan:
Augment existing tests in column_family_test
options_test
db_test
Reviewers: igor, IslamAbdelRahman, sdong, anthony
Reviewed By: anthony
Subscribers: dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D49095
2015-11-12 15:52:43 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
Status GetColumnFamilyOptionsFromMapInternal(
|
|
|
|
const ColumnFamilyOptions& base_options,
|
|
|
|
const std::unordered_map<std::string, std::string>& opts_map,
|
|
|
|
ColumnFamilyOptions* new_options, bool input_strings_escaped,
|
2017-06-14 01:55:08 +02:00
|
|
|
std::vector<std::string>* unsupported_options_names,
|
|
|
|
bool ignore_unknown_options) {
|
2014-09-17 21:46:32 +02:00
|
|
|
assert(new_options);
|
|
|
|
*new_options = base_options;
|
Add OptionsUtil::LoadOptionsFromFile() API
Summary:
This patch adds OptionsUtil::LoadOptionsFromFile() and
OptionsUtil::LoadLatestOptionsFromDB(), which allow developers
to construct DBOptions and ColumnFamilyOptions from a RocksDB
options file. Note that most pointer-typed options such as
merge_operator will not be constructed.
With this API, developers no longer need to remember all the
options in order to reopen an existing rocksdb instance like
the following:
DBOptions db_options;
std::vector<std::string> cf_names;
std::vector<ColumnFamilyOptions> cf_opts;
// Load primitive-typed options from an existing DB
OptionsUtil::LoadLatestOptionsFromDB(
dbname, &db_options, &cf_names, &cf_opts);
// Initialize necessary pointer-typed options
cf_opts[0].merge_operator.reset(new MyMergeOperator());
...
// Construct the vector of ColumnFamilyDescriptor
std::vector<ColumnFamilyDescriptor> cf_descs;
for (size_t i = 0; i < cf_opts.size(); ++i) {
cf_descs.emplace_back(cf_names[i], cf_opts[i]);
}
// Open the DB
DB* db = nullptr;
std::vector<ColumnFamilyHandle*> cf_handles;
auto s = DB::Open(db_options, dbname, cf_descs,
&handles, &db);
Test Plan:
Augment existing tests in column_family_test
options_test
db_test
Reviewers: igor, IslamAbdelRahman, sdong, anthony
Reviewed By: anthony
Subscribers: dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D49095
2015-11-12 15:52:43 +01:00
|
|
|
if (unsupported_options_names) {
|
|
|
|
unsupported_options_names->clear();
|
|
|
|
}
|
2014-10-10 19:00:12 +02:00
|
|
|
for (const auto& o : opts_map) {
|
Add OptionsUtil::LoadOptionsFromFile() API
Summary:
This patch adds OptionsUtil::LoadOptionsFromFile() and
OptionsUtil::LoadLatestOptionsFromDB(), which allow developers
to construct DBOptions and ColumnFamilyOptions from a RocksDB
options file. Note that most pointer-typed options such as
merge_operator will not be constructed.
With this API, developers no longer need to remember all the
options in order to reopen an existing rocksdb instance like
the following:
DBOptions db_options;
std::vector<std::string> cf_names;
std::vector<ColumnFamilyOptions> cf_opts;
// Load primitive-typed options from an existing DB
OptionsUtil::LoadLatestOptionsFromDB(
dbname, &db_options, &cf_names, &cf_opts);
// Initialize necessary pointer-typed options
cf_opts[0].merge_operator.reset(new MyMergeOperator());
...
// Construct the vector of ColumnFamilyDescriptor
std::vector<ColumnFamilyDescriptor> cf_descs;
for (size_t i = 0; i < cf_opts.size(); ++i) {
cf_descs.emplace_back(cf_names[i], cf_opts[i]);
}
// Open the DB
DB* db = nullptr;
std::vector<ColumnFamilyHandle*> cf_handles;
auto s = DB::Open(db_options, dbname, cf_descs,
&handles, &db);
Test Plan:
Augment existing tests in column_family_test
options_test
db_test
Reviewers: igor, IslamAbdelRahman, sdong, anthony
Reviewed By: anthony
Subscribers: dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D49095
2015-11-12 15:52:43 +01:00
|
|
|
auto s = ParseColumnFamilyOption(o.first, o.second, new_options,
|
|
|
|
input_strings_escaped);
|
|
|
|
if (!s.ok()) {
|
|
|
|
if (s.IsNotSupported()) {
|
|
|
|
// If the deserialization of the specified option is not supported
|
|
|
|
// and an output vector of unsupported_options is provided, then
|
|
|
|
// we log the name of the unsupported option and proceed.
|
|
|
|
if (unsupported_options_names != nullptr) {
|
|
|
|
unsupported_options_names->push_back(o.first);
|
|
|
|
}
|
|
|
|
// Note that we still return Status::OK in such case to maintain
|
|
|
|
// the backward compatibility in the old public API defined in
|
|
|
|
// rocksdb/convenience.h
|
2017-06-14 01:55:08 +02:00
|
|
|
} else if (s.IsInvalidArgument() && ignore_unknown_options) {
|
|
|
|
continue;
|
Add OptionsUtil::LoadOptionsFromFile() API
Summary:
This patch adds OptionsUtil::LoadOptionsFromFile() and
OptionsUtil::LoadLatestOptionsFromDB(), which allow developers
to construct DBOptions and ColumnFamilyOptions from a RocksDB
options file. Note that most pointer-typed options such as
merge_operator will not be constructed.
With this API, developers no longer need to remember all the
options in order to reopen an existing rocksdb instance like
the following:
DBOptions db_options;
std::vector<std::string> cf_names;
std::vector<ColumnFamilyOptions> cf_opts;
// Load primitive-typed options from an existing DB
OptionsUtil::LoadLatestOptionsFromDB(
dbname, &db_options, &cf_names, &cf_opts);
// Initialize necessary pointer-typed options
cf_opts[0].merge_operator.reset(new MyMergeOperator());
...
// Construct the vector of ColumnFamilyDescriptor
std::vector<ColumnFamilyDescriptor> cf_descs;
for (size_t i = 0; i < cf_opts.size(); ++i) {
cf_descs.emplace_back(cf_names[i], cf_opts[i]);
}
// Open the DB
DB* db = nullptr;
std::vector<ColumnFamilyHandle*> cf_handles;
auto s = DB::Open(db_options, dbname, cf_descs,
&handles, &db);
Test Plan:
Augment existing tests in column_family_test
options_test
db_test
Reviewers: igor, IslamAbdelRahman, sdong, anthony
Reviewed By: anthony
Subscribers: dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D49095
2015-11-12 15:52:43 +01:00
|
|
|
} else {
|
2016-08-11 23:54:29 +02:00
|
|
|
// Restore "new_options" to the default "base_options".
|
|
|
|
*new_options = base_options;
|
Add OptionsUtil::LoadOptionsFromFile() API
Summary:
This patch adds OptionsUtil::LoadOptionsFromFile() and
OptionsUtil::LoadLatestOptionsFromDB(), which allow developers
to construct DBOptions and ColumnFamilyOptions from a RocksDB
options file. Note that most pointer-typed options such as
merge_operator will not be constructed.
With this API, developers no longer need to remember all the
options in order to reopen an existing rocksdb instance like
the following:
DBOptions db_options;
std::vector<std::string> cf_names;
std::vector<ColumnFamilyOptions> cf_opts;
// Load primitive-typed options from an existing DB
OptionsUtil::LoadLatestOptionsFromDB(
dbname, &db_options, &cf_names, &cf_opts);
// Initialize necessary pointer-typed options
cf_opts[0].merge_operator.reset(new MyMergeOperator());
...
// Construct the vector of ColumnFamilyDescriptor
std::vector<ColumnFamilyDescriptor> cf_descs;
for (size_t i = 0; i < cf_opts.size(); ++i) {
cf_descs.emplace_back(cf_names[i], cf_opts[i]);
}
// Open the DB
DB* db = nullptr;
std::vector<ColumnFamilyHandle*> cf_handles;
auto s = DB::Open(db_options, dbname, cf_descs,
&handles, &db);
Test Plan:
Augment existing tests in column_family_test
options_test
db_test
Reviewers: igor, IslamAbdelRahman, sdong, anthony
Reviewed By: anthony
Subscribers: dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D49095
2015-11-12 15:52:43 +01:00
|
|
|
return s;
|
2015-10-03 00:35:32 +02:00
|
|
|
}
|
2014-10-10 19:00:12 +02:00
|
|
|
}
|
|
|
|
}
|
2014-12-22 22:18:57 +01:00
|
|
|
return Status::OK();
|
2014-10-10 19:00:12 +02:00
|
|
|
}
|
|
|
|
|
2014-12-22 22:18:57 +01:00
|
|
|
Status GetColumnFamilyOptionsFromString(
|
2014-10-10 19:00:12 +02:00
|
|
|
const ColumnFamilyOptions& base_options,
|
|
|
|
const std::string& opts_str,
|
|
|
|
ColumnFamilyOptions* new_options) {
|
|
|
|
std::unordered_map<std::string, std::string> opts_map;
|
2014-12-22 22:18:57 +01:00
|
|
|
Status s = StringToMap(opts_str, &opts_map);
|
|
|
|
if (!s.ok()) {
|
2016-08-11 23:54:29 +02:00
|
|
|
*new_options = base_options;
|
2014-12-22 22:18:57 +01:00
|
|
|
return s;
|
2014-10-10 19:00:12 +02:00
|
|
|
}
|
|
|
|
return GetColumnFamilyOptionsFromMap(base_options, opts_map, new_options);
|
|
|
|
}
|
|
|
|
|
2014-12-22 22:18:57 +01:00
|
|
|
Status GetDBOptionsFromMap(
|
2014-10-10 19:00:12 +02:00
|
|
|
const DBOptions& base_options,
|
|
|
|
const std::unordered_map<std::string, std::string>& opts_map,
|
2017-06-14 01:55:08 +02:00
|
|
|
DBOptions* new_options, bool input_strings_escaped,
|
|
|
|
bool ignore_unknown_options) {
|
|
|
|
return GetDBOptionsFromMapInternal(base_options, opts_map, new_options,
|
|
|
|
input_strings_escaped, nullptr,
|
|
|
|
ignore_unknown_options);
|
Add OptionsUtil::LoadOptionsFromFile() API
Summary:
This patch adds OptionsUtil::LoadOptionsFromFile() and
OptionsUtil::LoadLatestOptionsFromDB(), which allow developers
to construct DBOptions and ColumnFamilyOptions from a RocksDB
options file. Note that most pointer-typed options such as
merge_operator will not be constructed.
With this API, developers no longer need to remember all the
options in order to reopen an existing rocksdb instance like
the following:
DBOptions db_options;
std::vector<std::string> cf_names;
std::vector<ColumnFamilyOptions> cf_opts;
// Load primitive-typed options from an existing DB
OptionsUtil::LoadLatestOptionsFromDB(
dbname, &db_options, &cf_names, &cf_opts);
// Initialize necessary pointer-typed options
cf_opts[0].merge_operator.reset(new MyMergeOperator());
...
// Construct the vector of ColumnFamilyDescriptor
std::vector<ColumnFamilyDescriptor> cf_descs;
for (size_t i = 0; i < cf_opts.size(); ++i) {
cf_descs.emplace_back(cf_names[i], cf_opts[i]);
}
// Open the DB
DB* db = nullptr;
std::vector<ColumnFamilyHandle*> cf_handles;
auto s = DB::Open(db_options, dbname, cf_descs,
&handles, &db);
Test Plan:
Augment existing tests in column_family_test
options_test
db_test
Reviewers: igor, IslamAbdelRahman, sdong, anthony
Reviewed By: anthony
Subscribers: dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D49095
2015-11-12 15:52:43 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
Status GetDBOptionsFromMapInternal(
|
|
|
|
const DBOptions& base_options,
|
|
|
|
const std::unordered_map<std::string, std::string>& opts_map,
|
|
|
|
DBOptions* new_options, bool input_strings_escaped,
|
2017-06-14 01:55:08 +02:00
|
|
|
std::vector<std::string>* unsupported_options_names,
|
|
|
|
bool ignore_unknown_options) {
|
2014-10-10 19:00:12 +02:00
|
|
|
assert(new_options);
|
|
|
|
*new_options = base_options;
|
Add OptionsUtil::LoadOptionsFromFile() API
Summary:
This patch adds OptionsUtil::LoadOptionsFromFile() and
OptionsUtil::LoadLatestOptionsFromDB(), which allow developers
to construct DBOptions and ColumnFamilyOptions from a RocksDB
options file. Note that most pointer-typed options such as
merge_operator will not be constructed.
With this API, developers no longer need to remember all the
options in order to reopen an existing rocksdb instance like
the following:
DBOptions db_options;
std::vector<std::string> cf_names;
std::vector<ColumnFamilyOptions> cf_opts;
// Load primitive-typed options from an existing DB
OptionsUtil::LoadLatestOptionsFromDB(
dbname, &db_options, &cf_names, &cf_opts);
// Initialize necessary pointer-typed options
cf_opts[0].merge_operator.reset(new MyMergeOperator());
...
// Construct the vector of ColumnFamilyDescriptor
std::vector<ColumnFamilyDescriptor> cf_descs;
for (size_t i = 0; i < cf_opts.size(); ++i) {
cf_descs.emplace_back(cf_names[i], cf_opts[i]);
}
// Open the DB
DB* db = nullptr;
std::vector<ColumnFamilyHandle*> cf_handles;
auto s = DB::Open(db_options, dbname, cf_descs,
&handles, &db);
Test Plan:
Augment existing tests in column_family_test
options_test
db_test
Reviewers: igor, IslamAbdelRahman, sdong, anthony
Reviewed By: anthony
Subscribers: dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D49095
2015-11-12 15:52:43 +01:00
|
|
|
if (unsupported_options_names) {
|
|
|
|
unsupported_options_names->clear();
|
|
|
|
}
|
2014-10-10 19:00:12 +02:00
|
|
|
for (const auto& o : opts_map) {
|
Add OptionsUtil::LoadOptionsFromFile() API
Summary:
This patch adds OptionsUtil::LoadOptionsFromFile() and
OptionsUtil::LoadLatestOptionsFromDB(), which allow developers
to construct DBOptions and ColumnFamilyOptions from a RocksDB
options file. Note that most pointer-typed options such as
merge_operator will not be constructed.
With this API, developers no longer need to remember all the
options in order to reopen an existing rocksdb instance like
the following:
DBOptions db_options;
std::vector<std::string> cf_names;
std::vector<ColumnFamilyOptions> cf_opts;
// Load primitive-typed options from an existing DB
OptionsUtil::LoadLatestOptionsFromDB(
dbname, &db_options, &cf_names, &cf_opts);
// Initialize necessary pointer-typed options
cf_opts[0].merge_operator.reset(new MyMergeOperator());
...
// Construct the vector of ColumnFamilyDescriptor
std::vector<ColumnFamilyDescriptor> cf_descs;
for (size_t i = 0; i < cf_opts.size(); ++i) {
cf_descs.emplace_back(cf_names[i], cf_opts[i]);
}
// Open the DB
DB* db = nullptr;
std::vector<ColumnFamilyHandle*> cf_handles;
auto s = DB::Open(db_options, dbname, cf_descs,
&handles, &db);
Test Plan:
Augment existing tests in column_family_test
options_test
db_test
Reviewers: igor, IslamAbdelRahman, sdong, anthony
Reviewed By: anthony
Subscribers: dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D49095
2015-11-12 15:52:43 +01:00
|
|
|
auto s = ParseDBOption(o.first, o.second,
|
|
|
|
new_options, input_strings_escaped);
|
|
|
|
if (!s.ok()) {
|
|
|
|
if (s.IsNotSupported()) {
|
|
|
|
// If the deserialization of the specified option is not supported
|
|
|
|
// and an output vector of unsupported_options is provided, then
|
|
|
|
// we log the name of the unsupported option and proceed.
|
|
|
|
if (unsupported_options_names != nullptr) {
|
|
|
|
unsupported_options_names->push_back(o.first);
|
|
|
|
}
|
|
|
|
// Note that we still return Status::OK in such case to maintain
|
|
|
|
// the backward compatibility in the old public API defined in
|
|
|
|
// rocksdb/convenience.h
|
2017-06-14 01:55:08 +02:00
|
|
|
} else if (s.IsInvalidArgument() && ignore_unknown_options) {
|
|
|
|
continue;
|
Add OptionsUtil::LoadOptionsFromFile() API
Summary:
This patch adds OptionsUtil::LoadOptionsFromFile() and
OptionsUtil::LoadLatestOptionsFromDB(), which allow developers
to construct DBOptions and ColumnFamilyOptions from a RocksDB
options file. Note that most pointer-typed options such as
merge_operator will not be constructed.
With this API, developers no longer need to remember all the
options in order to reopen an existing rocksdb instance like
the following:
DBOptions db_options;
std::vector<std::string> cf_names;
std::vector<ColumnFamilyOptions> cf_opts;
// Load primitive-typed options from an existing DB
OptionsUtil::LoadLatestOptionsFromDB(
dbname, &db_options, &cf_names, &cf_opts);
// Initialize necessary pointer-typed options
cf_opts[0].merge_operator.reset(new MyMergeOperator());
...
// Construct the vector of ColumnFamilyDescriptor
std::vector<ColumnFamilyDescriptor> cf_descs;
for (size_t i = 0; i < cf_opts.size(); ++i) {
cf_descs.emplace_back(cf_names[i], cf_opts[i]);
}
// Open the DB
DB* db = nullptr;
std::vector<ColumnFamilyHandle*> cf_handles;
auto s = DB::Open(db_options, dbname, cf_descs,
&handles, &db);
Test Plan:
Augment existing tests in column_family_test
options_test
db_test
Reviewers: igor, IslamAbdelRahman, sdong, anthony
Reviewed By: anthony
Subscribers: dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D49095
2015-11-12 15:52:43 +01:00
|
|
|
} else {
|
2016-08-11 23:54:29 +02:00
|
|
|
// Restore "new_options" to the default "base_options".
|
|
|
|
*new_options = base_options;
|
Add OptionsUtil::LoadOptionsFromFile() API
Summary:
This patch adds OptionsUtil::LoadOptionsFromFile() and
OptionsUtil::LoadLatestOptionsFromDB(), which allow developers
to construct DBOptions and ColumnFamilyOptions from a RocksDB
options file. Note that most pointer-typed options such as
merge_operator will not be constructed.
With this API, developers no longer need to remember all the
options in order to reopen an existing rocksdb instance like
the following:
DBOptions db_options;
std::vector<std::string> cf_names;
std::vector<ColumnFamilyOptions> cf_opts;
// Load primitive-typed options from an existing DB
OptionsUtil::LoadLatestOptionsFromDB(
dbname, &db_options, &cf_names, &cf_opts);
// Initialize necessary pointer-typed options
cf_opts[0].merge_operator.reset(new MyMergeOperator());
...
// Construct the vector of ColumnFamilyDescriptor
std::vector<ColumnFamilyDescriptor> cf_descs;
for (size_t i = 0; i < cf_opts.size(); ++i) {
cf_descs.emplace_back(cf_names[i], cf_opts[i]);
}
// Open the DB
DB* db = nullptr;
std::vector<ColumnFamilyHandle*> cf_handles;
auto s = DB::Open(db_options, dbname, cf_descs,
&handles, &db);
Test Plan:
Augment existing tests in column_family_test
options_test
db_test
Reviewers: igor, IslamAbdelRahman, sdong, anthony
Reviewed By: anthony
Subscribers: dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D49095
2015-11-12 15:52:43 +01:00
|
|
|
return s;
|
|
|
|
}
|
2014-09-17 21:46:32 +02:00
|
|
|
}
|
|
|
|
}
|
2014-12-22 22:18:57 +01:00
|
|
|
return Status::OK();
|
2014-09-17 21:46:32 +02:00
|
|
|
}
|
|
|
|
|
2014-12-22 22:18:57 +01:00
|
|
|
Status GetDBOptionsFromString(
|
2014-10-10 19:00:12 +02:00
|
|
|
const DBOptions& base_options,
|
|
|
|
const std::string& opts_str,
|
|
|
|
DBOptions* new_options) {
|
|
|
|
std::unordered_map<std::string, std::string> opts_map;
|
2014-12-22 22:18:57 +01:00
|
|
|
Status s = StringToMap(opts_str, &opts_map);
|
|
|
|
if (!s.ok()) {
|
2016-08-11 23:54:29 +02:00
|
|
|
*new_options = base_options;
|
2014-12-22 22:18:57 +01:00
|
|
|
return s;
|
2014-10-10 19:00:12 +02:00
|
|
|
}
|
|
|
|
return GetDBOptionsFromMap(base_options, opts_map, new_options);
|
|
|
|
}
|
|
|
|
|
2015-02-20 04:11:14 +01:00
|
|
|
Status GetOptionsFromString(const Options& base_options,
|
|
|
|
const std::string& opts_str, Options* new_options) {
|
|
|
|
std::unordered_map<std::string, std::string> opts_map;
|
|
|
|
Status s = StringToMap(opts_str, &opts_map);
|
|
|
|
if (!s.ok()) {
|
|
|
|
return s;
|
|
|
|
}
|
|
|
|
DBOptions new_db_options(base_options);
|
|
|
|
ColumnFamilyOptions new_cf_options(base_options);
|
|
|
|
for (const auto& o : opts_map) {
|
Add OptionsUtil::LoadOptionsFromFile() API
Summary:
This patch adds OptionsUtil::LoadOptionsFromFile() and
OptionsUtil::LoadLatestOptionsFromDB(), which allow developers
to construct DBOptions and ColumnFamilyOptions from a RocksDB
options file. Note that most pointer-typed options such as
merge_operator will not be constructed.
With this API, developers no longer need to remember all the
options in order to reopen an existing rocksdb instance like
the following:
DBOptions db_options;
std::vector<std::string> cf_names;
std::vector<ColumnFamilyOptions> cf_opts;
// Load primitive-typed options from an existing DB
OptionsUtil::LoadLatestOptionsFromDB(
dbname, &db_options, &cf_names, &cf_opts);
// Initialize necessary pointer-typed options
cf_opts[0].merge_operator.reset(new MyMergeOperator());
...
// Construct the vector of ColumnFamilyDescriptor
std::vector<ColumnFamilyDescriptor> cf_descs;
for (size_t i = 0; i < cf_opts.size(); ++i) {
cf_descs.emplace_back(cf_names[i], cf_opts[i]);
}
// Open the DB
DB* db = nullptr;
std::vector<ColumnFamilyHandle*> cf_handles;
auto s = DB::Open(db_options, dbname, cf_descs,
&handles, &db);
Test Plan:
Augment existing tests in column_family_test
options_test
db_test
Reviewers: igor, IslamAbdelRahman, sdong, anthony
Reviewed By: anthony
Subscribers: dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D49095
2015-11-12 15:52:43 +01:00
|
|
|
if (ParseDBOption(o.first, o.second, &new_db_options).ok()) {
|
|
|
|
} else if (ParseColumnFamilyOption(
|
|
|
|
o.first, o.second, &new_cf_options).ok()) {
|
2015-02-20 04:11:14 +01:00
|
|
|
} else {
|
|
|
|
return Status::InvalidArgument("Can't parse option " + o.first);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
*new_options = Options(new_db_options, new_cf_options);
|
|
|
|
return Status::OK();
|
|
|
|
}
|
|
|
|
|
2015-10-11 21:17:42 +02:00
|
|
|
Status GetTableFactoryFromMap(
|
|
|
|
const std::string& factory_name,
|
|
|
|
const std::unordered_map<std::string, std::string>& opt_map,
|
2017-06-14 01:55:08 +02:00
|
|
|
std::shared_ptr<TableFactory>* table_factory, bool ignore_unknown_options) {
|
2015-10-11 21:17:42 +02:00
|
|
|
Status s;
|
|
|
|
if (factory_name == BlockBasedTableFactory().Name()) {
|
|
|
|
BlockBasedTableOptions bbt_opt;
|
|
|
|
s = GetBlockBasedTableOptionsFromMap(BlockBasedTableOptions(), opt_map,
|
2017-06-14 01:55:08 +02:00
|
|
|
&bbt_opt,
|
|
|
|
true, /* input_strings_escaped */
|
|
|
|
ignore_unknown_options);
|
2015-10-11 21:17:42 +02:00
|
|
|
if (!s.ok()) {
|
|
|
|
return s;
|
|
|
|
}
|
|
|
|
table_factory->reset(new BlockBasedTableFactory(bbt_opt));
|
|
|
|
return Status::OK();
|
2015-10-28 18:46:01 +01:00
|
|
|
} else if (factory_name == PlainTableFactory().Name()) {
|
|
|
|
PlainTableOptions pt_opt;
|
2015-11-16 21:56:21 +01:00
|
|
|
s = GetPlainTableOptionsFromMap(PlainTableOptions(), opt_map, &pt_opt,
|
2017-06-14 01:55:08 +02:00
|
|
|
true, /* input_strings_escaped */
|
|
|
|
ignore_unknown_options);
|
2015-10-28 18:46:01 +01:00
|
|
|
if (!s.ok()) {
|
|
|
|
return s;
|
|
|
|
}
|
|
|
|
table_factory->reset(new PlainTableFactory(pt_opt));
|
|
|
|
return Status::OK();
|
2015-10-11 21:17:42 +02:00
|
|
|
}
|
|
|
|
// Return OK for not supported table factories as TableFactory
|
|
|
|
// Deserialization is optional.
|
|
|
|
table_factory->reset();
|
|
|
|
return Status::OK();
|
|
|
|
}
|
|
|
|
|
2017-11-18 02:02:13 +01:00
|
|
|
std::unordered_map<std::string, OptionTypeInfo>
|
|
|
|
OptionsHelper::db_options_type_info = {
|
|
|
|
/*
|
|
|
|
// not yet supported
|
|
|
|
std::shared_ptr<Cache> row_cache;
|
|
|
|
std::shared_ptr<DeleteScheduler> delete_scheduler;
|
|
|
|
std::shared_ptr<Logger> info_log;
|
|
|
|
std::shared_ptr<RateLimiter> rate_limiter;
|
|
|
|
std::shared_ptr<Statistics> statistics;
|
|
|
|
std::vector<DbPath> db_paths;
|
|
|
|
std::vector<std::shared_ptr<EventListener>> listeners;
|
|
|
|
*/
|
|
|
|
{"advise_random_on_open",
|
|
|
|
{offsetof(struct DBOptions, advise_random_on_open),
|
|
|
|
OptionType::kBoolean, OptionVerificationType::kNormal, false, 0}},
|
|
|
|
{"allow_mmap_reads",
|
|
|
|
{offsetof(struct DBOptions, allow_mmap_reads), OptionType::kBoolean,
|
|
|
|
OptionVerificationType::kNormal, false, 0}},
|
|
|
|
{"allow_fallocate",
|
|
|
|
{offsetof(struct DBOptions, allow_fallocate), OptionType::kBoolean,
|
|
|
|
OptionVerificationType::kNormal, false, 0}},
|
|
|
|
{"allow_mmap_writes",
|
|
|
|
{offsetof(struct DBOptions, allow_mmap_writes), OptionType::kBoolean,
|
|
|
|
OptionVerificationType::kNormal, false, 0}},
|
|
|
|
{"use_direct_reads",
|
|
|
|
{offsetof(struct DBOptions, use_direct_reads), OptionType::kBoolean,
|
|
|
|
OptionVerificationType::kNormal, false, 0}},
|
|
|
|
{"use_direct_writes",
|
|
|
|
{0, OptionType::kBoolean, OptionVerificationType::kDeprecated, false,
|
|
|
|
0}},
|
|
|
|
{"use_direct_io_for_flush_and_compaction",
|
|
|
|
{offsetof(struct DBOptions, use_direct_io_for_flush_and_compaction),
|
|
|
|
OptionType::kBoolean, OptionVerificationType::kNormal, false, 0}},
|
|
|
|
{"allow_2pc",
|
|
|
|
{offsetof(struct DBOptions, allow_2pc), OptionType::kBoolean,
|
|
|
|
OptionVerificationType::kNormal, false, 0}},
|
|
|
|
{"allow_os_buffer",
|
|
|
|
{0, OptionType::kBoolean, OptionVerificationType::kDeprecated, true,
|
|
|
|
0}},
|
|
|
|
{"create_if_missing",
|
|
|
|
{offsetof(struct DBOptions, create_if_missing), OptionType::kBoolean,
|
|
|
|
OptionVerificationType::kNormal, false, 0}},
|
|
|
|
{"create_missing_column_families",
|
|
|
|
{offsetof(struct DBOptions, create_missing_column_families),
|
|
|
|
OptionType::kBoolean, OptionVerificationType::kNormal, false, 0}},
|
|
|
|
{"disableDataSync",
|
|
|
|
{0, OptionType::kBoolean, OptionVerificationType::kDeprecated, false,
|
|
|
|
0}},
|
|
|
|
{"disable_data_sync", // for compatibility
|
|
|
|
{0, OptionType::kBoolean, OptionVerificationType::kDeprecated, false,
|
|
|
|
0}},
|
|
|
|
{"enable_thread_tracking",
|
|
|
|
{offsetof(struct DBOptions, enable_thread_tracking),
|
|
|
|
OptionType::kBoolean, OptionVerificationType::kNormal, false, 0}},
|
|
|
|
{"error_if_exists",
|
|
|
|
{offsetof(struct DBOptions, error_if_exists), OptionType::kBoolean,
|
|
|
|
OptionVerificationType::kNormal, false, 0}},
|
|
|
|
{"is_fd_close_on_exec",
|
|
|
|
{offsetof(struct DBOptions, is_fd_close_on_exec), OptionType::kBoolean,
|
|
|
|
OptionVerificationType::kNormal, false, 0}},
|
|
|
|
{"paranoid_checks",
|
|
|
|
{offsetof(struct DBOptions, paranoid_checks), OptionType::kBoolean,
|
|
|
|
OptionVerificationType::kNormal, false, 0}},
|
|
|
|
{"skip_log_error_on_recovery",
|
|
|
|
{offsetof(struct DBOptions, skip_log_error_on_recovery),
|
|
|
|
OptionType::kBoolean, OptionVerificationType::kNormal, false, 0}},
|
|
|
|
{"skip_stats_update_on_db_open",
|
|
|
|
{offsetof(struct DBOptions, skip_stats_update_on_db_open),
|
|
|
|
OptionType::kBoolean, OptionVerificationType::kNormal, false, 0}},
|
|
|
|
{"new_table_reader_for_compaction_inputs",
|
|
|
|
{offsetof(struct DBOptions, new_table_reader_for_compaction_inputs),
|
|
|
|
OptionType::kBoolean, OptionVerificationType::kNormal, false, 0}},
|
|
|
|
{"compaction_readahead_size",
|
|
|
|
{offsetof(struct DBOptions, compaction_readahead_size),
|
|
|
|
OptionType::kSizeT, OptionVerificationType::kNormal, true,
|
|
|
|
offsetof(struct MutableDBOptions, compaction_readahead_size)}},
|
|
|
|
{"random_access_max_buffer_size",
|
|
|
|
{offsetof(struct DBOptions, random_access_max_buffer_size),
|
|
|
|
OptionType::kSizeT, OptionVerificationType::kNormal, false, 0}},
|
|
|
|
{"use_adaptive_mutex",
|
|
|
|
{offsetof(struct DBOptions, use_adaptive_mutex), OptionType::kBoolean,
|
|
|
|
OptionVerificationType::kNormal, false, 0}},
|
|
|
|
{"use_fsync",
|
|
|
|
{offsetof(struct DBOptions, use_fsync), OptionType::kBoolean,
|
|
|
|
OptionVerificationType::kNormal, false, 0}},
|
|
|
|
{"max_background_jobs",
|
|
|
|
{offsetof(struct DBOptions, max_background_jobs), OptionType::kInt,
|
|
|
|
OptionVerificationType::kNormal, true,
|
|
|
|
offsetof(struct MutableDBOptions, max_background_jobs)}},
|
|
|
|
{"max_background_compactions",
|
|
|
|
{offsetof(struct DBOptions, max_background_compactions),
|
|
|
|
OptionType::kInt, OptionVerificationType::kNormal, true,
|
|
|
|
offsetof(struct MutableDBOptions, max_background_compactions)}},
|
|
|
|
{"base_background_compactions",
|
|
|
|
{offsetof(struct DBOptions, base_background_compactions),
|
|
|
|
OptionType::kInt, OptionVerificationType::kNormal, true,
|
|
|
|
offsetof(struct MutableDBOptions, base_background_compactions)}},
|
|
|
|
{"max_background_flushes",
|
|
|
|
{offsetof(struct DBOptions, max_background_flushes), OptionType::kInt,
|
|
|
|
OptionVerificationType::kNormal, false, 0}},
|
|
|
|
{"max_file_opening_threads",
|
|
|
|
{offsetof(struct DBOptions, max_file_opening_threads),
|
|
|
|
OptionType::kInt, OptionVerificationType::kNormal, false, 0}},
|
|
|
|
{"max_open_files",
|
|
|
|
{offsetof(struct DBOptions, max_open_files), OptionType::kInt,
|
|
|
|
OptionVerificationType::kNormal, true,
|
|
|
|
offsetof(struct MutableDBOptions, max_open_files)}},
|
|
|
|
{"table_cache_numshardbits",
|
|
|
|
{offsetof(struct DBOptions, table_cache_numshardbits),
|
|
|
|
OptionType::kInt, OptionVerificationType::kNormal, false, 0}},
|
|
|
|
{"db_write_buffer_size",
|
|
|
|
{offsetof(struct DBOptions, db_write_buffer_size), OptionType::kSizeT,
|
|
|
|
OptionVerificationType::kNormal, false, 0}},
|
|
|
|
{"keep_log_file_num",
|
|
|
|
{offsetof(struct DBOptions, keep_log_file_num), OptionType::kSizeT,
|
|
|
|
OptionVerificationType::kNormal, false, 0}},
|
|
|
|
{"recycle_log_file_num",
|
|
|
|
{offsetof(struct DBOptions, recycle_log_file_num), OptionType::kSizeT,
|
|
|
|
OptionVerificationType::kNormal, false, 0}},
|
|
|
|
{"log_file_time_to_roll",
|
|
|
|
{offsetof(struct DBOptions, log_file_time_to_roll), OptionType::kSizeT,
|
|
|
|
OptionVerificationType::kNormal, false, 0}},
|
|
|
|
{"manifest_preallocation_size",
|
|
|
|
{offsetof(struct DBOptions, manifest_preallocation_size),
|
|
|
|
OptionType::kSizeT, OptionVerificationType::kNormal, false, 0}},
|
|
|
|
{"max_log_file_size",
|
|
|
|
{offsetof(struct DBOptions, max_log_file_size), OptionType::kSizeT,
|
|
|
|
OptionVerificationType::kNormal, false, 0}},
|
|
|
|
{"db_log_dir",
|
|
|
|
{offsetof(struct DBOptions, db_log_dir), OptionType::kString,
|
|
|
|
OptionVerificationType::kNormal, false, 0}},
|
|
|
|
{"wal_dir",
|
|
|
|
{offsetof(struct DBOptions, wal_dir), OptionType::kString,
|
|
|
|
OptionVerificationType::kNormal, false, 0}},
|
|
|
|
{"max_subcompactions",
|
|
|
|
{offsetof(struct DBOptions, max_subcompactions), OptionType::kUInt32T,
|
|
|
|
OptionVerificationType::kNormal, false, 0}},
|
|
|
|
{"WAL_size_limit_MB",
|
|
|
|
{offsetof(struct DBOptions, WAL_size_limit_MB), OptionType::kUInt64T,
|
|
|
|
OptionVerificationType::kNormal, false, 0}},
|
|
|
|
{"WAL_ttl_seconds",
|
|
|
|
{offsetof(struct DBOptions, WAL_ttl_seconds), OptionType::kUInt64T,
|
|
|
|
OptionVerificationType::kNormal, false, 0}},
|
|
|
|
{"bytes_per_sync",
|
|
|
|
{offsetof(struct DBOptions, bytes_per_sync), OptionType::kUInt64T,
|
|
|
|
OptionVerificationType::kNormal, true,
|
|
|
|
offsetof(struct MutableDBOptions, bytes_per_sync)}},
|
|
|
|
{"delayed_write_rate",
|
|
|
|
{offsetof(struct DBOptions, delayed_write_rate), OptionType::kUInt64T,
|
|
|
|
OptionVerificationType::kNormal, true,
|
|
|
|
offsetof(struct MutableDBOptions, delayed_write_rate)}},
|
|
|
|
{"delete_obsolete_files_period_micros",
|
|
|
|
{offsetof(struct DBOptions, delete_obsolete_files_period_micros),
|
|
|
|
OptionType::kUInt64T, OptionVerificationType::kNormal, true,
|
|
|
|
offsetof(struct MutableDBOptions,
|
|
|
|
delete_obsolete_files_period_micros)}},
|
|
|
|
{"max_manifest_file_size",
|
|
|
|
{offsetof(struct DBOptions, max_manifest_file_size),
|
|
|
|
OptionType::kUInt64T, OptionVerificationType::kNormal, false, 0}},
|
|
|
|
{"max_total_wal_size",
|
|
|
|
{offsetof(struct DBOptions, max_total_wal_size), OptionType::kUInt64T,
|
|
|
|
OptionVerificationType::kNormal, true,
|
|
|
|
offsetof(struct MutableDBOptions, max_total_wal_size)}},
|
|
|
|
{"wal_bytes_per_sync",
|
|
|
|
{offsetof(struct DBOptions, wal_bytes_per_sync), OptionType::kUInt64T,
|
|
|
|
OptionVerificationType::kNormal, true,
|
|
|
|
offsetof(struct MutableDBOptions, wal_bytes_per_sync)}},
|
Optionally wait on bytes_per_sync to smooth I/O (#5183)
Summary:
The existing implementation does not guarantee bytes reach disk every `bytes_per_sync` when writing SST files, or every `wal_bytes_per_sync` when writing WALs. This can cause confusing behavior for users who enable this feature to avoid large syncs during flush and compaction, but then end up hitting them anyways.
My understanding of the existing behavior is we used `sync_file_range` with `SYNC_FILE_RANGE_WRITE` to submit ranges for async writeback, such that we could continue processing the next range of bytes while that I/O is happening. I believe we can preserve that benefit while also limiting how far the processing can get ahead of the I/O, which prevents huge syncs from happening when the file finishes.
Consider this `sync_file_range` usage: `sync_file_range(fd_, 0, static_cast<off_t>(offset + nbytes), SYNC_FILE_RANGE_WAIT_BEFORE | SYNC_FILE_RANGE_WRITE)`. Expanding the range to start at 0 and adding the `SYNC_FILE_RANGE_WAIT_BEFORE` flag causes any pending writeback (like from a previous call to `sync_file_range`) to finish before it proceeds to submit the latest `nbytes` for writeback. The latest `nbytes` are still written back asynchronously, unless processing exceeds I/O speed, in which case the following `sync_file_range` will need to wait on it.
There is a second change in this PR to use `fdatasync` when `sync_file_range` is unavailable (determined statically) or has some known problem with the underlying filesystem (determined dynamically).
The above two changes only apply when the user enables a new option, `strict_bytes_per_sync`.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5183
Differential Revision: D14953553
Pulled By: siying
fbshipit-source-id: 445c3862e019fb7b470f9c7f314fc231b62706e9
2019-04-22 20:48:45 +02:00
|
|
|
{"strict_bytes_per_sync",
|
2019-04-25 20:31:58 +02:00
|
|
|
{offsetof(struct DBOptions, strict_bytes_per_sync),
|
|
|
|
OptionType::kBoolean, OptionVerificationType::kNormal, true,
|
Optionally wait on bytes_per_sync to smooth I/O (#5183)
Summary:
The existing implementation does not guarantee bytes reach disk every `bytes_per_sync` when writing SST files, or every `wal_bytes_per_sync` when writing WALs. This can cause confusing behavior for users who enable this feature to avoid large syncs during flush and compaction, but then end up hitting them anyways.
My understanding of the existing behavior is we used `sync_file_range` with `SYNC_FILE_RANGE_WRITE` to submit ranges for async writeback, such that we could continue processing the next range of bytes while that I/O is happening. I believe we can preserve that benefit while also limiting how far the processing can get ahead of the I/O, which prevents huge syncs from happening when the file finishes.
Consider this `sync_file_range` usage: `sync_file_range(fd_, 0, static_cast<off_t>(offset + nbytes), SYNC_FILE_RANGE_WAIT_BEFORE | SYNC_FILE_RANGE_WRITE)`. Expanding the range to start at 0 and adding the `SYNC_FILE_RANGE_WAIT_BEFORE` flag causes any pending writeback (like from a previous call to `sync_file_range`) to finish before it proceeds to submit the latest `nbytes` for writeback. The latest `nbytes` are still written back asynchronously, unless processing exceeds I/O speed, in which case the following `sync_file_range` will need to wait on it.
There is a second change in this PR to use `fdatasync` when `sync_file_range` is unavailable (determined statically) or has some known problem with the underlying filesystem (determined dynamically).
The above two changes only apply when the user enables a new option, `strict_bytes_per_sync`.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5183
Differential Revision: D14953553
Pulled By: siying
fbshipit-source-id: 445c3862e019fb7b470f9c7f314fc231b62706e9
2019-04-22 20:48:45 +02:00
|
|
|
offsetof(struct MutableDBOptions, strict_bytes_per_sync)}},
|
2017-11-18 02:02:13 +01:00
|
|
|
{"stats_dump_period_sec",
|
|
|
|
{offsetof(struct DBOptions, stats_dump_period_sec), OptionType::kUInt,
|
|
|
|
OptionVerificationType::kNormal, true,
|
|
|
|
offsetof(struct MutableDBOptions, stats_dump_period_sec)}},
|
2019-02-21 00:46:59 +01:00
|
|
|
{"stats_persist_period_sec",
|
|
|
|
{offsetof(struct DBOptions, stats_persist_period_sec),
|
|
|
|
OptionType::kUInt, OptionVerificationType::kNormal, true,
|
|
|
|
offsetof(struct MutableDBOptions, stats_persist_period_sec)}},
|
2019-06-18 00:17:43 +02:00
|
|
|
{"persist_stats_to_disk",
|
|
|
|
{offsetof(struct DBOptions, persist_stats_to_disk),
|
|
|
|
OptionType::kBoolean, OptionVerificationType::kNormal, false,
|
|
|
|
offsetof(struct ImmutableDBOptions, persist_stats_to_disk)}},
|
2019-02-21 00:46:59 +01:00
|
|
|
{"stats_history_buffer_size",
|
|
|
|
{offsetof(struct DBOptions, stats_history_buffer_size),
|
|
|
|
OptionType::kSizeT, OptionVerificationType::kNormal, true,
|
|
|
|
offsetof(struct MutableDBOptions, stats_history_buffer_size)}},
|
2017-11-18 02:02:13 +01:00
|
|
|
{"fail_if_options_file_error",
|
|
|
|
{offsetof(struct DBOptions, fail_if_options_file_error),
|
|
|
|
OptionType::kBoolean, OptionVerificationType::kNormal, false, 0}},
|
|
|
|
{"enable_pipelined_write",
|
|
|
|
{offsetof(struct DBOptions, enable_pipelined_write),
|
|
|
|
OptionType::kBoolean, OptionVerificationType::kNormal, false, 0}},
|
2019-05-14 02:43:47 +02:00
|
|
|
{"unordered_write",
|
|
|
|
{offsetof(struct DBOptions, unordered_write), OptionType::kBoolean,
|
|
|
|
OptionVerificationType::kNormal, false, 0}},
|
2017-11-18 02:02:13 +01:00
|
|
|
{"allow_concurrent_memtable_write",
|
|
|
|
{offsetof(struct DBOptions, allow_concurrent_memtable_write),
|
|
|
|
OptionType::kBoolean, OptionVerificationType::kNormal, false, 0}},
|
|
|
|
{"wal_recovery_mode",
|
|
|
|
{offsetof(struct DBOptions, wal_recovery_mode),
|
|
|
|
OptionType::kWALRecoveryMode, OptionVerificationType::kNormal, false,
|
|
|
|
0}},
|
|
|
|
{"enable_write_thread_adaptive_yield",
|
|
|
|
{offsetof(struct DBOptions, enable_write_thread_adaptive_yield),
|
|
|
|
OptionType::kBoolean, OptionVerificationType::kNormal, false, 0}},
|
|
|
|
{"write_thread_slow_yield_usec",
|
|
|
|
{offsetof(struct DBOptions, write_thread_slow_yield_usec),
|
|
|
|
OptionType::kUInt64T, OptionVerificationType::kNormal, false, 0}},
|
|
|
|
{"write_thread_max_yield_usec",
|
|
|
|
{offsetof(struct DBOptions, write_thread_max_yield_usec),
|
|
|
|
OptionType::kUInt64T, OptionVerificationType::kNormal, false, 0}},
|
|
|
|
{"access_hint_on_compaction_start",
|
|
|
|
{offsetof(struct DBOptions, access_hint_on_compaction_start),
|
|
|
|
OptionType::kAccessHint, OptionVerificationType::kNormal, false, 0}},
|
|
|
|
{"info_log_level",
|
|
|
|
{offsetof(struct DBOptions, info_log_level), OptionType::kInfoLogLevel,
|
|
|
|
OptionVerificationType::kNormal, false, 0}},
|
|
|
|
{"dump_malloc_stats",
|
|
|
|
{offsetof(struct DBOptions, dump_malloc_stats), OptionType::kBoolean,
|
|
|
|
OptionVerificationType::kNormal, false, 0}},
|
|
|
|
{"avoid_flush_during_recovery",
|
|
|
|
{offsetof(struct DBOptions, avoid_flush_during_recovery),
|
|
|
|
OptionType::kBoolean, OptionVerificationType::kNormal, false, 0}},
|
|
|
|
{"avoid_flush_during_shutdown",
|
|
|
|
{offsetof(struct DBOptions, avoid_flush_during_shutdown),
|
|
|
|
OptionType::kBoolean, OptionVerificationType::kNormal, true,
|
|
|
|
offsetof(struct MutableDBOptions, avoid_flush_during_shutdown)}},
|
|
|
|
{"writable_file_max_buffer_size",
|
|
|
|
{offsetof(struct DBOptions, writable_file_max_buffer_size),
|
|
|
|
OptionType::kSizeT, OptionVerificationType::kNormal, true,
|
|
|
|
offsetof(struct MutableDBOptions, writable_file_max_buffer_size)}},
|
|
|
|
{"allow_ingest_behind",
|
|
|
|
{offsetof(struct DBOptions, allow_ingest_behind), OptionType::kBoolean,
|
|
|
|
OptionVerificationType::kNormal, false,
|
|
|
|
offsetof(struct ImmutableDBOptions, allow_ingest_behind)}},
|
|
|
|
{"preserve_deletes",
|
|
|
|
{offsetof(struct DBOptions, preserve_deletes), OptionType::kBoolean,
|
|
|
|
OptionVerificationType::kNormal, false,
|
|
|
|
offsetof(struct ImmutableDBOptions, preserve_deletes)}},
|
|
|
|
{"concurrent_prepare", // Deprecated by two_write_queues
|
|
|
|
{0, OptionType::kBoolean, OptionVerificationType::kDeprecated, false,
|
|
|
|
0}},
|
|
|
|
{"two_write_queues",
|
|
|
|
{offsetof(struct DBOptions, two_write_queues), OptionType::kBoolean,
|
|
|
|
OptionVerificationType::kNormal, false,
|
|
|
|
offsetof(struct ImmutableDBOptions, two_write_queues)}},
|
|
|
|
{"manual_wal_flush",
|
|
|
|
{offsetof(struct DBOptions, manual_wal_flush), OptionType::kBoolean,
|
|
|
|
OptionVerificationType::kNormal, false,
|
|
|
|
offsetof(struct ImmutableDBOptions, manual_wal_flush)}},
|
|
|
|
{"seq_per_batch",
|
|
|
|
{0, OptionType::kBoolean, OptionVerificationType::kDeprecated, false,
|
2018-10-27 00:06:44 +02:00
|
|
|
0}},
|
|
|
|
{"atomic_flush",
|
|
|
|
{offsetof(struct DBOptions, atomic_flush), OptionType::kBoolean,
|
|
|
|
OptionVerificationType::kNormal, false,
|
2019-04-02 02:07:38 +02:00
|
|
|
offsetof(struct ImmutableDBOptions, atomic_flush)}},
|
|
|
|
{"avoid_unnecessary_blocking_io",
|
|
|
|
{offsetof(struct DBOptions, avoid_unnecessary_blocking_io),
|
|
|
|
OptionType::kBoolean, OptionVerificationType::kNormal, false,
|
2019-04-25 20:31:58 +02:00
|
|
|
offsetof(struct ImmutableDBOptions, avoid_unnecessary_blocking_io)}},
|
2019-07-19 20:54:38 +02:00
|
|
|
{"log_readahead_size",
|
|
|
|
{offsetof(struct DBOptions, log_readahead_size), OptionType::kSizeT,
|
|
|
|
OptionVerificationType::kNormal, false, 0}},
|
2019-04-25 20:31:58 +02:00
|
|
|
};
|
2017-11-18 02:02:13 +01:00
|
|
|
|
|
|
|
std::unordered_map<std::string, BlockBasedTableOptions::IndexType>
|
|
|
|
OptionsHelper::block_base_table_index_type_string_map = {
|
|
|
|
{"kBinarySearch", BlockBasedTableOptions::IndexType::kBinarySearch},
|
|
|
|
{"kHashSearch", BlockBasedTableOptions::IndexType::kHashSearch},
|
|
|
|
{"kTwoLevelIndexSearch",
|
Add an option to put first key of each sst block in the index (#5289)
Summary:
The first key is used to defer reading the data block until this file gets to the top of merging iterator's heap. For short range scans, most files never make it to the top of the heap, so this change can reduce read amplification by a lot sometimes.
Consider the following workload. There are a few data streams (we'll be calling them "logs"), each stream consisting of a sequence of blobs (we'll be calling them "records"). Each record is identified by log ID and a sequence number within the log. RocksDB key is concatenation of log ID and sequence number (big endian). Reads are mostly relatively short range scans, each within a single log. Writes are mostly sequential for each log, but writes to different logs are randomly interleaved. Compactions are disabled; instead, when we accumulate a few tens of sst files, we create a new column family and start writing to it.
So, a typical sst file consists of a few ranges of blocks, each range corresponding to one log ID (we use FlushBlockPolicy to cut blocks at log boundaries). A typical read would go like this. First, iterator Seek() reads one block from each sst file. Then a series of Next()s move through one sst file (since writes to each log are mostly sequential) until the subiterator reaches the end of this log in this sst file; then Next() switches to the next sst file and reads sequentially from that, and so on. Often a range scan will only return records from a small number of blocks in small number of sst files; in this case, the cost of initial Seek() reading one block from each file may be bigger than the cost of reading the actually useful blocks.
Neither iterate_upper_bound nor bloom filters can prevent reading one block from each file in Seek(). But this PR can: if the index contains first key from each block, we don't have to read the block until this block actually makes it to the top of merging iterator's heap, so for short range scans we won't read any blocks from most of the sst files.
This PR does the deferred block loading inside value() call. This is not ideal: there's no good way to report an IO error from inside value(). As discussed with siying offline, it would probably be better to change InternalIterator's interface to explicitly fetch deferred value and get status. I'll do it in a separate PR.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5289
Differential Revision: D15256423
Pulled By: al13n321
fbshipit-source-id: 750e4c39ce88e8d41662f701cf6275d9388ba46a
2019-06-25 05:50:35 +02:00
|
|
|
BlockBasedTableOptions::IndexType::kTwoLevelIndexSearch},
|
|
|
|
{"kBinarySearchWithFirstKey",
|
|
|
|
BlockBasedTableOptions::IndexType::kBinarySearchWithFirstKey}};
|
2017-11-18 02:02:13 +01:00
|
|
|
|
2018-07-28 00:35:41 +02:00
|
|
|
std::unordered_map<std::string, BlockBasedTableOptions::DataBlockIndexType>
|
|
|
|
OptionsHelper::block_base_table_data_block_index_type_string_map = {
|
|
|
|
{"kDataBlockBinarySearch",
|
|
|
|
BlockBasedTableOptions::DataBlockIndexType::kDataBlockBinarySearch},
|
2018-08-15 23:27:47 +02:00
|
|
|
{"kDataBlockBinaryAndHash",
|
|
|
|
BlockBasedTableOptions::DataBlockIndexType::kDataBlockBinaryAndHash}};
|
2018-07-28 00:35:41 +02:00
|
|
|
|
2019-04-22 17:17:45 +02:00
|
|
|
std::unordered_map<std::string, BlockBasedTableOptions::IndexShorteningMode>
|
|
|
|
OptionsHelper::block_base_table_index_shortening_mode_string_map = {
|
2019-05-04 02:26:20 +02:00
|
|
|
{"kNoShortening",
|
|
|
|
BlockBasedTableOptions::IndexShorteningMode::kNoShortening},
|
|
|
|
{"kShortenSeparators",
|
|
|
|
BlockBasedTableOptions::IndexShorteningMode::kShortenSeparators},
|
|
|
|
{"kShortenSeparatorsAndSuccessor",
|
|
|
|
BlockBasedTableOptions::IndexShorteningMode::
|
|
|
|
kShortenSeparatorsAndSuccessor}};
|
2019-04-22 17:17:45 +02:00
|
|
|
|
2017-11-18 02:02:13 +01:00
|
|
|
std::unordered_map<std::string, EncodingType>
|
|
|
|
OptionsHelper::encoding_type_string_map = {{"kPlain", kPlain},
|
|
|
|
{"kPrefix", kPrefix}};
|
|
|
|
|
|
|
|
std::unordered_map<std::string, CompactionStyle>
|
|
|
|
OptionsHelper::compaction_style_string_map = {
|
|
|
|
{"kCompactionStyleLevel", kCompactionStyleLevel},
|
|
|
|
{"kCompactionStyleUniversal", kCompactionStyleUniversal},
|
|
|
|
{"kCompactionStyleFIFO", kCompactionStyleFIFO},
|
|
|
|
{"kCompactionStyleNone", kCompactionStyleNone}};
|
|
|
|
|
|
|
|
std::unordered_map<std::string, CompactionPri>
|
|
|
|
OptionsHelper::compaction_pri_string_map = {
|
|
|
|
{"kByCompensatedSize", kByCompensatedSize},
|
|
|
|
{"kOldestLargestSeqFirst", kOldestLargestSeqFirst},
|
|
|
|
{"kOldestSmallestSeqFirst", kOldestSmallestSeqFirst},
|
|
|
|
{"kMinOverlappingRatio", kMinOverlappingRatio}};
|
|
|
|
|
|
|
|
std::unordered_map<std::string, WALRecoveryMode>
|
|
|
|
OptionsHelper::wal_recovery_mode_string_map = {
|
|
|
|
{"kTolerateCorruptedTailRecords",
|
|
|
|
WALRecoveryMode::kTolerateCorruptedTailRecords},
|
|
|
|
{"kAbsoluteConsistency", WALRecoveryMode::kAbsoluteConsistency},
|
|
|
|
{"kPointInTimeRecovery", WALRecoveryMode::kPointInTimeRecovery},
|
|
|
|
{"kSkipAnyCorruptedRecords",
|
|
|
|
WALRecoveryMode::kSkipAnyCorruptedRecords}};
|
|
|
|
|
|
|
|
std::unordered_map<std::string, DBOptions::AccessHint>
|
|
|
|
OptionsHelper::access_hint_string_map = {
|
|
|
|
{"NONE", DBOptions::AccessHint::NONE},
|
|
|
|
{"NORMAL", DBOptions::AccessHint::NORMAL},
|
|
|
|
{"SEQUENTIAL", DBOptions::AccessHint::SEQUENTIAL},
|
|
|
|
{"WILLNEED", DBOptions::AccessHint::WILLNEED}};
|
|
|
|
|
|
|
|
std::unordered_map<std::string, InfoLogLevel>
|
|
|
|
OptionsHelper::info_log_level_string_map = {
|
|
|
|
{"DEBUG_LEVEL", InfoLogLevel::DEBUG_LEVEL},
|
|
|
|
{"INFO_LEVEL", InfoLogLevel::INFO_LEVEL},
|
|
|
|
{"WARN_LEVEL", InfoLogLevel::WARN_LEVEL},
|
|
|
|
{"ERROR_LEVEL", InfoLogLevel::ERROR_LEVEL},
|
|
|
|
{"FATAL_LEVEL", InfoLogLevel::FATAL_LEVEL},
|
|
|
|
{"HEADER_LEVEL", InfoLogLevel::HEADER_LEVEL}};
|
|
|
|
|
|
|
|
ColumnFamilyOptions OptionsHelper::dummy_cf_options;
|
|
|
|
CompactionOptionsFIFO OptionsHelper::dummy_comp_options;
|
2017-11-28 19:35:17 +01:00
|
|
|
LRUCacheOptions OptionsHelper::dummy_lru_cache_options;
|
2017-12-11 22:12:12 +01:00
|
|
|
CompactionOptionsUniversal OptionsHelper::dummy_comp_options_universal;
|
2017-11-18 02:02:13 +01:00
|
|
|
|
|
|
|
// offset_of is used to get the offset of a class data member
|
|
|
|
// ex: offset_of(&ColumnFamilyOptions::num_levels)
|
|
|
|
// This call will return the offset of num_levels in ColumnFamilyOptions class
|
|
|
|
//
|
|
|
|
// This is the same as offsetof() but allow us to work with non standard-layout
|
|
|
|
// classes and structures
|
|
|
|
// refs:
|
|
|
|
// http://en.cppreference.com/w/cpp/concept/StandardLayoutType
|
|
|
|
// https://gist.github.com/graphitemaster/494f21190bb2c63c5516
|
|
|
|
template <typename T1>
|
|
|
|
int offset_of(T1 ColumnFamilyOptions::*member) {
|
|
|
|
return int(size_t(&(OptionsHelper::dummy_cf_options.*member)) -
|
|
|
|
size_t(&OptionsHelper::dummy_cf_options));
|
|
|
|
}
|
|
|
|
template <typename T1>
|
|
|
|
int offset_of(T1 AdvancedColumnFamilyOptions::*member) {
|
|
|
|
return int(size_t(&(OptionsHelper::dummy_cf_options.*member)) -
|
|
|
|
size_t(&OptionsHelper::dummy_cf_options));
|
|
|
|
}
|
|
|
|
template <typename T1>
|
|
|
|
int offset_of(T1 CompactionOptionsFIFO::*member) {
|
|
|
|
return int(size_t(&(OptionsHelper::dummy_comp_options.*member)) -
|
|
|
|
size_t(&OptionsHelper::dummy_comp_options));
|
|
|
|
}
|
2017-11-28 19:35:17 +01:00
|
|
|
template <typename T1>
|
|
|
|
int offset_of(T1 LRUCacheOptions::*member) {
|
|
|
|
return int(size_t(&(OptionsHelper::dummy_lru_cache_options.*member)) -
|
|
|
|
size_t(&OptionsHelper::dummy_lru_cache_options));
|
|
|
|
}
|
2017-12-11 22:12:12 +01:00
|
|
|
template <typename T1>
|
|
|
|
int offset_of(T1 CompactionOptionsUniversal::*member) {
|
|
|
|
return int(size_t(&(OptionsHelper::dummy_comp_options_universal.*member)) -
|
|
|
|
size_t(&OptionsHelper::dummy_comp_options_universal));
|
|
|
|
}
|
2017-11-18 02:02:13 +01:00
|
|
|
|
|
|
|
std::unordered_map<std::string, OptionTypeInfo>
|
|
|
|
OptionsHelper::cf_options_type_info = {
|
|
|
|
/* not yet supported
|
|
|
|
CompressionOptions compression_opts;
|
|
|
|
TablePropertiesCollectorFactories table_properties_collector_factories;
|
|
|
|
typedef std::vector<std::shared_ptr<TablePropertiesCollectorFactory>>
|
|
|
|
TablePropertiesCollectorFactories;
|
|
|
|
UpdateStatus (*inplace_callback)(char* existing_value,
|
|
|
|
uint34_t* existing_value_size,
|
|
|
|
Slice delta_value,
|
|
|
|
std::string* merged_value);
|
2018-04-06 04:49:06 +02:00
|
|
|
std::vector<DbPath> cf_paths;
|
2017-11-18 02:02:13 +01:00
|
|
|
*/
|
|
|
|
{"report_bg_io_stats",
|
|
|
|
{offset_of(&ColumnFamilyOptions::report_bg_io_stats),
|
|
|
|
OptionType::kBoolean, OptionVerificationType::kNormal, true,
|
|
|
|
offsetof(struct MutableCFOptions, report_bg_io_stats)}},
|
|
|
|
{"compaction_measure_io_stats",
|
|
|
|
{0, OptionType::kBoolean, OptionVerificationType::kDeprecated, false,
|
|
|
|
0}},
|
|
|
|
{"disable_auto_compactions",
|
|
|
|
{offset_of(&ColumnFamilyOptions::disable_auto_compactions),
|
|
|
|
OptionType::kBoolean, OptionVerificationType::kNormal, true,
|
|
|
|
offsetof(struct MutableCFOptions, disable_auto_compactions)}},
|
|
|
|
{"filter_deletes",
|
|
|
|
{0, OptionType::kBoolean, OptionVerificationType::kDeprecated, true,
|
|
|
|
0}},
|
|
|
|
{"inplace_update_support",
|
|
|
|
{offset_of(&ColumnFamilyOptions::inplace_update_support),
|
|
|
|
OptionType::kBoolean, OptionVerificationType::kNormal, false, 0}},
|
|
|
|
{"level_compaction_dynamic_level_bytes",
|
|
|
|
{offset_of(&ColumnFamilyOptions::level_compaction_dynamic_level_bytes),
|
|
|
|
OptionType::kBoolean, OptionVerificationType::kNormal, false, 0}},
|
|
|
|
{"optimize_filters_for_hits",
|
|
|
|
{offset_of(&ColumnFamilyOptions::optimize_filters_for_hits),
|
|
|
|
OptionType::kBoolean, OptionVerificationType::kNormal, false, 0}},
|
|
|
|
{"paranoid_file_checks",
|
|
|
|
{offset_of(&ColumnFamilyOptions::paranoid_file_checks),
|
|
|
|
OptionType::kBoolean, OptionVerificationType::kNormal, true,
|
|
|
|
offsetof(struct MutableCFOptions, paranoid_file_checks)}},
|
|
|
|
{"force_consistency_checks",
|
|
|
|
{offset_of(&ColumnFamilyOptions::force_consistency_checks),
|
|
|
|
OptionType::kBoolean, OptionVerificationType::kNormal, false, 0}},
|
|
|
|
{"purge_redundant_kvs_while_flush",
|
|
|
|
{offset_of(&ColumnFamilyOptions::purge_redundant_kvs_while_flush),
|
|
|
|
OptionType::kBoolean, OptionVerificationType::kDeprecated, false, 0}},
|
|
|
|
{"verify_checksums_in_compaction",
|
|
|
|
{0, OptionType::kBoolean, OptionVerificationType::kDeprecated, true,
|
|
|
|
0}},
|
|
|
|
{"soft_pending_compaction_bytes_limit",
|
|
|
|
{offset_of(&ColumnFamilyOptions::soft_pending_compaction_bytes_limit),
|
|
|
|
OptionType::kUInt64T, OptionVerificationType::kNormal, true,
|
|
|
|
offsetof(struct MutableCFOptions,
|
|
|
|
soft_pending_compaction_bytes_limit)}},
|
|
|
|
{"hard_pending_compaction_bytes_limit",
|
|
|
|
{offset_of(&ColumnFamilyOptions::hard_pending_compaction_bytes_limit),
|
|
|
|
OptionType::kUInt64T, OptionVerificationType::kNormal, true,
|
|
|
|
offsetof(struct MutableCFOptions,
|
|
|
|
hard_pending_compaction_bytes_limit)}},
|
|
|
|
{"hard_rate_limit",
|
|
|
|
{0, OptionType::kDouble, OptionVerificationType::kDeprecated, true,
|
|
|
|
0}},
|
|
|
|
{"soft_rate_limit",
|
|
|
|
{0, OptionType::kDouble, OptionVerificationType::kDeprecated, true,
|
|
|
|
0}},
|
|
|
|
{"max_compaction_bytes",
|
|
|
|
{offset_of(&ColumnFamilyOptions::max_compaction_bytes),
|
|
|
|
OptionType::kUInt64T, OptionVerificationType::kNormal, true,
|
|
|
|
offsetof(struct MutableCFOptions, max_compaction_bytes)}},
|
|
|
|
{"expanded_compaction_factor",
|
|
|
|
{0, OptionType::kInt, OptionVerificationType::kDeprecated, true, 0}},
|
|
|
|
{"level0_file_num_compaction_trigger",
|
|
|
|
{offset_of(&ColumnFamilyOptions::level0_file_num_compaction_trigger),
|
|
|
|
OptionType::kInt, OptionVerificationType::kNormal, true,
|
|
|
|
offsetof(struct MutableCFOptions,
|
|
|
|
level0_file_num_compaction_trigger)}},
|
|
|
|
{"level0_slowdown_writes_trigger",
|
|
|
|
{offset_of(&ColumnFamilyOptions::level0_slowdown_writes_trigger),
|
|
|
|
OptionType::kInt, OptionVerificationType::kNormal, true,
|
|
|
|
offsetof(struct MutableCFOptions, level0_slowdown_writes_trigger)}},
|
|
|
|
{"level0_stop_writes_trigger",
|
|
|
|
{offset_of(&ColumnFamilyOptions::level0_stop_writes_trigger),
|
|
|
|
OptionType::kInt, OptionVerificationType::kNormal, true,
|
|
|
|
offsetof(struct MutableCFOptions, level0_stop_writes_trigger)}},
|
|
|
|
{"max_grandparent_overlap_factor",
|
|
|
|
{0, OptionType::kInt, OptionVerificationType::kDeprecated, true, 0}},
|
|
|
|
{"max_mem_compaction_level",
|
|
|
|
{0, OptionType::kInt, OptionVerificationType::kDeprecated, false, 0}},
|
|
|
|
{"max_write_buffer_number",
|
|
|
|
{offset_of(&ColumnFamilyOptions::max_write_buffer_number),
|
|
|
|
OptionType::kInt, OptionVerificationType::kNormal, true,
|
|
|
|
offsetof(struct MutableCFOptions, max_write_buffer_number)}},
|
|
|
|
{"max_write_buffer_number_to_maintain",
|
|
|
|
{offset_of(&ColumnFamilyOptions::max_write_buffer_number_to_maintain),
|
|
|
|
OptionType::kInt, OptionVerificationType::kNormal, false, 0}},
|
Refactor trimming logic for immutable memtables (#5022)
Summary:
MyRocks currently sets `max_write_buffer_number_to_maintain` in order to maintain enough history for transaction conflict checking. The effectiveness of this approach depends on the size of memtables. When memtables are small, it may not keep enough history; when memtables are large, this may consume too much memory.
We are proposing a new way to configure memtable list history: by limiting the memory usage of immutable memtables. The new option is `max_write_buffer_size_to_maintain` and it will take precedence over the old `max_write_buffer_number_to_maintain` if they are both set to non-zero values. The new option accounts for the total memory usage of flushed immutable memtables and mutable memtable. When the total usage exceeds the limit, RocksDB may start dropping immutable memtables (which is also called trimming history), starting from the oldest one.
The semantics of the old option actually works both as an upper bound and lower bound. History trimming will start if number of immutable memtables exceeds the limit, but it will never go below (limit-1) due to history trimming.
In order the mimic the behavior with the new option, history trimming will stop if dropping the next immutable memtable causes the total memory usage go below the size limit. For example, assuming the size limit is set to 64MB, and there are 3 immutable memtables with sizes of 20, 30, 30. Although the total memory usage is 80MB > 64MB, dropping the oldest memtable will reduce the memory usage to 60MB < 64MB, so in this case no memtable will be dropped.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5022
Differential Revision: D14394062
Pulled By: miasantreble
fbshipit-source-id: 60457a509c6af89d0993f988c9b5c2aa9e45f5c5
2019-08-23 22:54:09 +02:00
|
|
|
{"max_write_buffer_size_to_maintain",
|
|
|
|
{offset_of(&ColumnFamilyOptions::max_write_buffer_size_to_maintain),
|
|
|
|
OptionType::kInt64T, OptionVerificationType::kNormal, false, 0}},
|
2017-11-18 02:02:13 +01:00
|
|
|
{"min_write_buffer_number_to_merge",
|
|
|
|
{offset_of(&ColumnFamilyOptions::min_write_buffer_number_to_merge),
|
|
|
|
OptionType::kInt, OptionVerificationType::kNormal, false, 0}},
|
|
|
|
{"num_levels",
|
|
|
|
{offset_of(&ColumnFamilyOptions::num_levels), OptionType::kInt,
|
|
|
|
OptionVerificationType::kNormal, false, 0}},
|
|
|
|
{"source_compaction_factor",
|
|
|
|
{0, OptionType::kInt, OptionVerificationType::kDeprecated, true, 0}},
|
|
|
|
{"target_file_size_multiplier",
|
|
|
|
{offset_of(&ColumnFamilyOptions::target_file_size_multiplier),
|
|
|
|
OptionType::kInt, OptionVerificationType::kNormal, true,
|
|
|
|
offsetof(struct MutableCFOptions, target_file_size_multiplier)}},
|
|
|
|
{"arena_block_size",
|
|
|
|
{offset_of(&ColumnFamilyOptions::arena_block_size), OptionType::kSizeT,
|
|
|
|
OptionVerificationType::kNormal, true,
|
|
|
|
offsetof(struct MutableCFOptions, arena_block_size)}},
|
|
|
|
{"inplace_update_num_locks",
|
|
|
|
{offset_of(&ColumnFamilyOptions::inplace_update_num_locks),
|
|
|
|
OptionType::kSizeT, OptionVerificationType::kNormal, true,
|
|
|
|
offsetof(struct MutableCFOptions, inplace_update_num_locks)}},
|
|
|
|
{"max_successive_merges",
|
|
|
|
{offset_of(&ColumnFamilyOptions::max_successive_merges),
|
|
|
|
OptionType::kSizeT, OptionVerificationType::kNormal, true,
|
|
|
|
offsetof(struct MutableCFOptions, max_successive_merges)}},
|
|
|
|
{"memtable_huge_page_size",
|
|
|
|
{offset_of(&ColumnFamilyOptions::memtable_huge_page_size),
|
|
|
|
OptionType::kSizeT, OptionVerificationType::kNormal, true,
|
|
|
|
offsetof(struct MutableCFOptions, memtable_huge_page_size)}},
|
|
|
|
{"memtable_prefix_bloom_huge_page_tlb_size",
|
|
|
|
{0, OptionType::kSizeT, OptionVerificationType::kDeprecated, true, 0}},
|
|
|
|
{"write_buffer_size",
|
|
|
|
{offset_of(&ColumnFamilyOptions::write_buffer_size),
|
|
|
|
OptionType::kSizeT, OptionVerificationType::kNormal, true,
|
|
|
|
offsetof(struct MutableCFOptions, write_buffer_size)}},
|
|
|
|
{"bloom_locality",
|
|
|
|
{offset_of(&ColumnFamilyOptions::bloom_locality), OptionType::kUInt32T,
|
|
|
|
OptionVerificationType::kNormal, false, 0}},
|
|
|
|
{"memtable_prefix_bloom_bits",
|
|
|
|
{0, OptionType::kUInt32T, OptionVerificationType::kDeprecated, true,
|
|
|
|
0}},
|
|
|
|
{"memtable_prefix_bloom_size_ratio",
|
|
|
|
{offset_of(&ColumnFamilyOptions::memtable_prefix_bloom_size_ratio),
|
|
|
|
OptionType::kDouble, OptionVerificationType::kNormal, true,
|
|
|
|
offsetof(struct MutableCFOptions, memtable_prefix_bloom_size_ratio)}},
|
|
|
|
{"memtable_prefix_bloom_probes",
|
|
|
|
{0, OptionType::kUInt32T, OptionVerificationType::kDeprecated, true,
|
|
|
|
0}},
|
2019-02-19 21:12:25 +01:00
|
|
|
{"memtable_whole_key_filtering",
|
|
|
|
{offset_of(&ColumnFamilyOptions::memtable_whole_key_filtering),
|
|
|
|
OptionType::kBoolean, OptionVerificationType::kNormal, true,
|
|
|
|
offsetof(struct MutableCFOptions, memtable_whole_key_filtering)}},
|
2017-11-18 02:02:13 +01:00
|
|
|
{"min_partial_merge_operands",
|
|
|
|
{0, OptionType::kUInt32T, OptionVerificationType::kDeprecated, true,
|
|
|
|
0}},
|
|
|
|
{"max_bytes_for_level_base",
|
|
|
|
{offset_of(&ColumnFamilyOptions::max_bytes_for_level_base),
|
|
|
|
OptionType::kUInt64T, OptionVerificationType::kNormal, true,
|
|
|
|
offsetof(struct MutableCFOptions, max_bytes_for_level_base)}},
|
2019-05-04 02:26:20 +02:00
|
|
|
{"snap_refresh_nanos",
|
|
|
|
{offset_of(&ColumnFamilyOptions::snap_refresh_nanos),
|
|
|
|
OptionType::kUInt64T, OptionVerificationType::kNormal, true,
|
|
|
|
offsetof(struct MutableCFOptions, snap_refresh_nanos)}},
|
2017-11-18 02:02:13 +01:00
|
|
|
{"max_bytes_for_level_multiplier",
|
|
|
|
{offset_of(&ColumnFamilyOptions::max_bytes_for_level_multiplier),
|
|
|
|
OptionType::kDouble, OptionVerificationType::kNormal, true,
|
|
|
|
offsetof(struct MutableCFOptions, max_bytes_for_level_multiplier)}},
|
|
|
|
{"max_bytes_for_level_multiplier_additional",
|
|
|
|
{offset_of(
|
|
|
|
&ColumnFamilyOptions::max_bytes_for_level_multiplier_additional),
|
|
|
|
OptionType::kVectorInt, OptionVerificationType::kNormal, true,
|
|
|
|
offsetof(struct MutableCFOptions,
|
|
|
|
max_bytes_for_level_multiplier_additional)}},
|
|
|
|
{"max_sequential_skip_in_iterations",
|
|
|
|
{offset_of(&ColumnFamilyOptions::max_sequential_skip_in_iterations),
|
|
|
|
OptionType::kUInt64T, OptionVerificationType::kNormal, true,
|
|
|
|
offsetof(struct MutableCFOptions,
|
|
|
|
max_sequential_skip_in_iterations)}},
|
|
|
|
{"target_file_size_base",
|
|
|
|
{offset_of(&ColumnFamilyOptions::target_file_size_base),
|
|
|
|
OptionType::kUInt64T, OptionVerificationType::kNormal, true,
|
|
|
|
offsetof(struct MutableCFOptions, target_file_size_base)}},
|
|
|
|
{"rate_limit_delay_max_milliseconds",
|
|
|
|
{0, OptionType::kUInt, OptionVerificationType::kDeprecated, false, 0}},
|
|
|
|
{"compression",
|
|
|
|
{offset_of(&ColumnFamilyOptions::compression),
|
|
|
|
OptionType::kCompressionType, OptionVerificationType::kNormal, true,
|
|
|
|
offsetof(struct MutableCFOptions, compression)}},
|
|
|
|
{"compression_per_level",
|
|
|
|
{offset_of(&ColumnFamilyOptions::compression_per_level),
|
|
|
|
OptionType::kVectorCompressionType, OptionVerificationType::kNormal,
|
|
|
|
false, 0}},
|
|
|
|
{"bottommost_compression",
|
|
|
|
{offset_of(&ColumnFamilyOptions::bottommost_compression),
|
|
|
|
OptionType::kCompressionType, OptionVerificationType::kNormal, false,
|
|
|
|
0}},
|
2019-03-26 22:15:26 +01:00
|
|
|
{kNameComparator,
|
2017-11-18 02:02:13 +01:00
|
|
|
{offset_of(&ColumnFamilyOptions::comparator), OptionType::kComparator,
|
|
|
|
OptionVerificationType::kByName, false, 0}},
|
|
|
|
{"prefix_extractor",
|
|
|
|
{offset_of(&ColumnFamilyOptions::prefix_extractor),
|
|
|
|
OptionType::kSliceTransform, OptionVerificationType::kByNameAllowNull,
|
2018-05-21 23:33:55 +02:00
|
|
|
true, offsetof(struct MutableCFOptions, prefix_extractor)}},
|
2017-11-18 02:02:13 +01:00
|
|
|
{"memtable_insert_with_hint_prefix_extractor",
|
|
|
|
{offset_of(
|
|
|
|
&ColumnFamilyOptions::memtable_insert_with_hint_prefix_extractor),
|
|
|
|
OptionType::kSliceTransform, OptionVerificationType::kByNameAllowNull,
|
|
|
|
false, 0}},
|
|
|
|
{"memtable_factory",
|
|
|
|
{offset_of(&ColumnFamilyOptions::memtable_factory),
|
|
|
|
OptionType::kMemTableRepFactory, OptionVerificationType::kByName,
|
|
|
|
false, 0}},
|
|
|
|
{"table_factory",
|
|
|
|
{offset_of(&ColumnFamilyOptions::table_factory),
|
|
|
|
OptionType::kTableFactory, OptionVerificationType::kByName, false,
|
|
|
|
0}},
|
|
|
|
{"compaction_filter",
|
|
|
|
{offset_of(&ColumnFamilyOptions::compaction_filter),
|
|
|
|
OptionType::kCompactionFilter, OptionVerificationType::kByName, false,
|
|
|
|
0}},
|
|
|
|
{"compaction_filter_factory",
|
|
|
|
{offset_of(&ColumnFamilyOptions::compaction_filter_factory),
|
|
|
|
OptionType::kCompactionFilterFactory, OptionVerificationType::kByName,
|
|
|
|
false, 0}},
|
2019-03-28 22:50:06 +01:00
|
|
|
{kNameMergeOperator,
|
2017-11-18 02:02:13 +01:00
|
|
|
{offset_of(&ColumnFamilyOptions::merge_operator),
|
|
|
|
OptionType::kMergeOperator,
|
|
|
|
OptionVerificationType::kByNameAllowFromNull, false, 0}},
|
|
|
|
{"compaction_style",
|
|
|
|
{offset_of(&ColumnFamilyOptions::compaction_style),
|
|
|
|
OptionType::kCompactionStyle, OptionVerificationType::kNormal, false,
|
|
|
|
0}},
|
|
|
|
{"compaction_pri",
|
|
|
|
{offset_of(&ColumnFamilyOptions::compaction_pri),
|
|
|
|
OptionType::kCompactionPri, OptionVerificationType::kNormal, false,
|
|
|
|
0}},
|
|
|
|
{"compaction_options_fifo",
|
|
|
|
{offset_of(&ColumnFamilyOptions::compaction_options_fifo),
|
|
|
|
OptionType::kCompactionOptionsFIFO, OptionVerificationType::kNormal,
|
2017-12-11 22:12:12 +01:00
|
|
|
true, offsetof(struct MutableCFOptions, compaction_options_fifo)}},
|
|
|
|
{"compaction_options_universal",
|
|
|
|
{offset_of(&ColumnFamilyOptions::compaction_options_universal),
|
|
|
|
OptionType::kCompactionOptionsUniversal,
|
|
|
|
OptionVerificationType::kNormal, true,
|
2018-04-03 06:57:28 +02:00
|
|
|
offsetof(struct MutableCFOptions, compaction_options_universal)}},
|
|
|
|
{"ttl",
|
|
|
|
{offset_of(&ColumnFamilyOptions::ttl), OptionType::kUInt64T,
|
2018-07-16 23:24:33 +02:00
|
|
|
OptionVerificationType::kNormal, true,
|
2019-03-18 20:07:35 +01:00
|
|
|
offsetof(struct MutableCFOptions, ttl)}},
|
Periodic Compactions (#5166)
Summary:
Introducing Periodic Compactions.
This feature allows all the files in a CF to be periodically compacted. It could help in catching any corruptions that could creep into the DB proactively as every file is constantly getting re-compacted. And also, of course, it helps to cleanup data older than certain threshold.
- Introduced a new option `periodic_compaction_time` to control how long a file can live without being compacted in a CF.
- This works across all levels.
- The files are put in the same level after going through the compaction. (Related files in the same level are picked up as `ExpandInputstoCleanCut` is used).
- Compaction filters, if any, are invoked as usual.
- A new table property, `file_creation_time`, is introduced to implement this feature. This property is set to the time at which the SST file was created (and that time is given by the underlying Env/OS).
This feature can be enabled on its own, or in conjunction with `ttl`. It is possible to set a different time threshold for the bottom level when used in conjunction with ttl. Since `ttl` works only on 0 to last but one levels, you could set `ttl` to, say, 1 day, and `periodic_compaction_time` to, say, 7 days. Since `ttl < periodic_compaction_time` all files in last but one levels keep getting picked up based on ttl, and almost never based on periodic_compaction_time. The files in the bottom level get picked up for compaction based on `periodic_compaction_time`.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/5166
Differential Revision: D14884441
Pulled By: sagar0
fbshipit-source-id: 408426cbacb409c06386a98632dcf90bfa1bda47
2019-04-11 04:24:25 +02:00
|
|
|
{"periodic_compaction_seconds",
|
|
|
|
{offset_of(&ColumnFamilyOptions::periodic_compaction_seconds),
|
|
|
|
OptionType::kUInt64T, OptionVerificationType::kNormal, true,
|
|
|
|
offsetof(struct MutableCFOptions, periodic_compaction_seconds)}},
|
2019-03-18 20:07:35 +01:00
|
|
|
{"sample_for_compression",
|
|
|
|
{offset_of(&ColumnFamilyOptions::sample_for_compression),
|
|
|
|
OptionType::kUInt64T, OptionVerificationType::kNormal, true,
|
|
|
|
offsetof(struct MutableCFOptions, sample_for_compression)}}};
|
2017-11-18 02:02:13 +01:00
|
|
|
|
|
|
|
std::unordered_map<std::string, OptionTypeInfo>
|
|
|
|
OptionsHelper::fifo_compaction_options_type_info = {
|
|
|
|
{"max_table_files_size",
|
|
|
|
{offset_of(&CompactionOptionsFIFO::max_table_files_size),
|
|
|
|
OptionType::kUInt64T, OptionVerificationType::kNormal, true,
|
|
|
|
offsetof(struct CompactionOptionsFIFO, max_table_files_size)}},
|
|
|
|
{"ttl",
|
2019-02-15 18:48:44 +01:00
|
|
|
{0, OptionType::kUInt64T,
|
|
|
|
OptionVerificationType::kDeprecated, false,
|
|
|
|
0}},
|
2017-11-18 02:02:13 +01:00
|
|
|
{"allow_compaction",
|
|
|
|
{offset_of(&CompactionOptionsFIFO::allow_compaction),
|
|
|
|
OptionType::kBoolean, OptionVerificationType::kNormal, true,
|
|
|
|
offsetof(struct CompactionOptionsFIFO, allow_compaction)}}};
|
|
|
|
|
2017-12-11 22:12:12 +01:00
|
|
|
std::unordered_map<std::string, OptionTypeInfo>
|
|
|
|
OptionsHelper::universal_compaction_options_type_info = {
|
|
|
|
{"size_ratio",
|
|
|
|
{offset_of(&CompactionOptionsUniversal::size_ratio), OptionType::kUInt,
|
|
|
|
OptionVerificationType::kNormal, true,
|
|
|
|
offsetof(class CompactionOptionsUniversal, size_ratio)}},
|
|
|
|
{"min_merge_width",
|
|
|
|
{offset_of(&CompactionOptionsUniversal::min_merge_width),
|
|
|
|
OptionType::kUInt, OptionVerificationType::kNormal, true,
|
|
|
|
offsetof(class CompactionOptionsUniversal, min_merge_width)}},
|
|
|
|
{"max_merge_width",
|
|
|
|
{offset_of(&CompactionOptionsUniversal::max_merge_width),
|
|
|
|
OptionType::kUInt, OptionVerificationType::kNormal, true,
|
|
|
|
offsetof(class CompactionOptionsUniversal, max_merge_width)}},
|
|
|
|
{"max_size_amplification_percent",
|
|
|
|
{offset_of(
|
|
|
|
&CompactionOptionsUniversal::max_size_amplification_percent),
|
|
|
|
OptionType::kUInt, OptionVerificationType::kNormal, true,
|
|
|
|
offsetof(class CompactionOptionsUniversal,
|
|
|
|
max_size_amplification_percent)}},
|
|
|
|
{"compression_size_percent",
|
|
|
|
{offset_of(&CompactionOptionsUniversal::compression_size_percent),
|
|
|
|
OptionType::kInt, OptionVerificationType::kNormal, true,
|
|
|
|
offsetof(class CompactionOptionsUniversal,
|
|
|
|
compression_size_percent)}},
|
|
|
|
{"stop_style",
|
|
|
|
{offset_of(&CompactionOptionsUniversal::stop_style),
|
|
|
|
OptionType::kCompactionStopStyle, OptionVerificationType::kNormal,
|
|
|
|
true, offsetof(class CompactionOptionsUniversal, stop_style)}},
|
|
|
|
{"allow_trivial_move",
|
|
|
|
{offset_of(&CompactionOptionsUniversal::allow_trivial_move),
|
|
|
|
OptionType::kBoolean, OptionVerificationType::kNormal, true,
|
|
|
|
offsetof(class CompactionOptionsUniversal, allow_trivial_move)}}};
|
|
|
|
|
|
|
|
std::unordered_map<std::string, CompactionStopStyle>
|
|
|
|
OptionsHelper::compaction_stop_style_string_map = {
|
|
|
|
{"kCompactionStopStyleSimilarSize", kCompactionStopStyleSimilarSize},
|
|
|
|
{"kCompactionStopStyleTotalSize", kCompactionStopStyleTotalSize}};
|
|
|
|
|
2017-11-28 19:35:17 +01:00
|
|
|
std::unordered_map<std::string, OptionTypeInfo>
|
|
|
|
OptionsHelper::lru_cache_options_type_info = {
|
2017-12-11 22:12:12 +01:00
|
|
|
{"capacity",
|
|
|
|
{offset_of(&LRUCacheOptions::capacity), OptionType::kSizeT,
|
|
|
|
OptionVerificationType::kNormal, true,
|
2017-11-28 19:35:17 +01:00
|
|
|
offsetof(struct LRUCacheOptions, capacity)}},
|
2017-12-11 22:12:12 +01:00
|
|
|
{"num_shard_bits",
|
|
|
|
{offset_of(&LRUCacheOptions::num_shard_bits), OptionType::kInt,
|
|
|
|
OptionVerificationType::kNormal, true,
|
2017-11-28 19:35:17 +01:00
|
|
|
offsetof(struct LRUCacheOptions, num_shard_bits)}},
|
|
|
|
{"strict_capacity_limit",
|
|
|
|
{offset_of(&LRUCacheOptions::strict_capacity_limit),
|
|
|
|
OptionType::kBoolean, OptionVerificationType::kNormal, true,
|
|
|
|
offsetof(struct LRUCacheOptions, strict_capacity_limit)}},
|
|
|
|
{"high_pri_pool_ratio",
|
2017-12-11 22:12:12 +01:00
|
|
|
{offset_of(&LRUCacheOptions::high_pri_pool_ratio), OptionType::kDouble,
|
|
|
|
OptionVerificationType::kNormal, true,
|
2017-11-28 19:35:17 +01:00
|
|
|
offsetof(struct LRUCacheOptions, high_pri_pool_ratio)}}};
|
|
|
|
|
RocksDB Options file format and its serialization / deserialization.
Summary:
This patch defines the format of RocksDB options file, which
follows the INI file format, and implements functions for its
serialization and deserialization. An example RocksDB options
file can be found in examples/rocksdb_option_file_example.ini.
A typical RocksDB options file has three sections, which are
Version, DBOptions, and more than one CFOptions. The RocksDB
options file in general follows the basic INI file format
with the following extensions / modifications:
* Escaped characters
We escaped the following characters:
- \n -- line feed - new line
- \r -- carriage return
- \\ -- backslash \
- \: -- colon symbol :
- \# -- hash tag #
* Comments
We support # style comments. Comments can appear at the ending
part of a line.
* Statements
A statement is of the form option_name = value.
Each statement contains a '=', where extra white-spaces
are supported. However, we don't support multi-lined statement.
Furthermore, each line can only contain at most one statement.
* Section
Sections are of the form [SecitonTitle "SectionArgument"],
where section argument is optional.
* List
We use colon-separated string to represent a list.
For instance, n1:n2:n3:n4 is a list containing four values.
Below is an example of a RocksDB options file:
[Version]
rocksdb_version=4.0.0
options_file_version=1.0
[DBOptions]
max_open_files=12345
max_background_flushes=301
[CFOptions "default"]
[CFOptions "the second column family"]
[CFOptions "the third column family"]
Test Plan: Added many tests in options_test.cc
Reviewers: igor, IslamAbdelRahman, sdong, anthony
Reviewed By: anthony
Subscribers: maykov, dhruba, leveldb
Differential Revision: https://reviews.facebook.net/D46059
2015-09-29 23:42:40 +02:00
|
|
|
#endif // !ROCKSDB_LITE
|
2016-09-15 07:10:28 +02:00
|
|
|
|
2014-09-17 21:46:32 +02:00
|
|
|
} // namespace rocksdb
|