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).
|
2013-12-05 22:09:13 +01:00
|
|
|
|
|
|
|
#include "rocksdb/table_properties.h"
|
2019-05-30 23:47:29 +02:00
|
|
|
|
2022-04-06 19:33:00 +02:00
|
|
|
#include "port/malloc.h"
|
2014-11-25 05:44:49 +01:00
|
|
|
#include "port/port.h"
|
2016-08-20 00:10:31 +02:00
|
|
|
#include "rocksdb/env.h"
|
Experimental support for SST unique IDs (#8990)
Summary:
* New public header unique_id.h and function GetUniqueIdFromTableProperties
which computes a universally unique identifier based on table properties
of table files from recent RocksDB versions.
* Generation of DB session IDs is refactored so that they are
guaranteed unique in the lifetime of a process running RocksDB.
(SemiStructuredUniqueIdGen, new test included.) Along with file numbers,
this enables SST unique IDs to be guaranteed unique among SSTs generated
in a single process, and "better than random" between processes.
See https://github.com/pdillinger/unique_id
* In addition to public API producing 'external' unique IDs, there is a function
for producing 'internal' unique IDs, with functions for converting between the
two. In short, the external ID is "safe" for things people might do with it, and
the internal ID enables more "power user" features for the future. Specifically,
the external ID goes through a hashing layer so that any subset of bits in the
external ID can be used as a hash of the full ID, while also preserving
uniqueness guarantees in the first 128 bits (bijective both on first 128 bits
and on full 192 bits).
Intended follow-up:
* Use the internal unique IDs in cache keys. (Avoid conflicts with https://github.com/facebook/rocksdb/issues/8912) (The file offset can be XORed into
the third 64-bit value of the unique ID.)
* Publish the external unique IDs in FileStorageInfo (https://github.com/facebook/rocksdb/issues/8968)
Pull Request resolved: https://github.com/facebook/rocksdb/pull/8990
Test Plan:
Unit tests added, and checking of unique ids in stress test.
NOTE in stress test we do not generate nearly enough files to thoroughly
stress uniqueness, but the test trims off pieces of the ID to check for
uniqueness so that we can infer (with some assumptions) stronger
properties in the aggregate.
Reviewed By: zhichao-cao, mrambacher
Differential Revision: D31582865
Pulled By: pdillinger
fbshipit-source-id: 1f620c4c86af9abe2a8d177b9ccf2ad2b9f48243
2021-10-19 08:28:28 +02:00
|
|
|
#include "rocksdb/unique_id.h"
|
2016-08-20 00:10:31 +02:00
|
|
|
#include "table/table_properties_internal.h"
|
Experimental support for SST unique IDs (#8990)
Summary:
* New public header unique_id.h and function GetUniqueIdFromTableProperties
which computes a universally unique identifier based on table properties
of table files from recent RocksDB versions.
* Generation of DB session IDs is refactored so that they are
guaranteed unique in the lifetime of a process running RocksDB.
(SemiStructuredUniqueIdGen, new test included.) Along with file numbers,
this enables SST unique IDs to be guaranteed unique among SSTs generated
in a single process, and "better than random" between processes.
See https://github.com/pdillinger/unique_id
* In addition to public API producing 'external' unique IDs, there is a function
for producing 'internal' unique IDs, with functions for converting between the
two. In short, the external ID is "safe" for things people might do with it, and
the internal ID enables more "power user" features for the future. Specifically,
the external ID goes through a hashing layer so that any subset of bits in the
external ID can be used as a hash of the full ID, while also preserving
uniqueness guarantees in the first 128 bits (bijective both on first 128 bits
and on full 192 bits).
Intended follow-up:
* Use the internal unique IDs in cache keys. (Avoid conflicts with https://github.com/facebook/rocksdb/issues/8912) (The file offset can be XORed into
the third 64-bit value of the unique ID.)
* Publish the external unique IDs in FileStorageInfo (https://github.com/facebook/rocksdb/issues/8968)
Pull Request resolved: https://github.com/facebook/rocksdb/pull/8990
Test Plan:
Unit tests added, and checking of unique ids in stress test.
NOTE in stress test we do not generate nearly enough files to thoroughly
stress uniqueness, but the test trims off pieces of the ID to check for
uniqueness so that we can infer (with some assumptions) stronger
properties in the aggregate.
Reviewed By: zhichao-cao, mrambacher
Differential Revision: D31582865
Pulled By: pdillinger
fbshipit-source-id: 1f620c4c86af9abe2a8d177b9ccf2ad2b9f48243
2021-10-19 08:28:28 +02:00
|
|
|
#include "table/unique_id_impl.h"
|
Improve / clean up meta block code & integrity (#9163)
Summary:
* Checksums are now checked on meta blocks unless specifically
suppressed or not applicable (e.g. plain table). (Was other way around.)
This means a number of cases that were not checking checksums now are,
including direct read TableProperties in Version::GetTableProperties
(fixed in meta_blocks ReadTableProperties), reading any block from
PersistentCache (fixed in BlockFetcher), read TableProperties in
SstFileDumper (ldb/sst_dump/BackupEngine) before table reader open,
maybe more.
* For that to work, I moved the global_seqno+TableProperties checksum
logic to the shared table/ code, because that is used by many utilies
such as SstFileDumper.
* Also for that to work, we have to know when we're dealing with a block
that has a checksum (trailer), so added that capability to Footer based
on magic number, and from there BlockFetcher.
* Knowledge of trailer presence has also fixed a problem where other
table formats were reading blocks including bytes for a non-existant
trailer--and awkwardly kind-of not using them, e.g. no shared code
checking checksums. (BlockFetcher compression type was populated
incorrectly.) Now we only read what is needed.
* Minimized code duplication and differing/incompatible/awkward
abstractions in meta_blocks.{cc,h} (e.g. SeekTo in metaindex block
without parsing block handle)
* Moved some meta block handling code from table_properties*.*
* Moved some code specific to block-based table from shared table/ code
to BlockBasedTable class. The checksum stuff means we can't completely
separate it, but things that don't need to be in shared table/ code
should not be.
* Use unique_ptr rather than raw ptr in more places. (Note: you can
std::move from unique_ptr to shared_ptr.)
Without enhancements to GetPropertiesOfAllTablesTest (see below),
net reduction of roughly 100 lines of code.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/9163
Test Plan:
existing tests and
* Enhanced DBTablePropertiesTest.GetPropertiesOfAllTablesTest to verify that
checksums are now checked on direct read of table properties by TableCache
(new test would fail before this change)
* Also enhanced DBTablePropertiesTest.GetPropertiesOfAllTablesTest to test
putting table properties under old meta name
* Also generally enhanced that same test to actually test what it was
supposed to be testing already, by kicking things out of table cache when
we don't want them there.
Reviewed By: ajkr, mrambacher
Differential Revision: D32514757
Pulled By: pdillinger
fbshipit-source-id: 507964b9311d186ae8d1131182290cbd97a99fa9
2021-11-18 20:42:12 +01:00
|
|
|
#include "util/random.h"
|
2014-11-25 05:44:49 +01:00
|
|
|
#include "util/string_util.h"
|
2013-12-05 22:09:13 +01:00
|
|
|
|
2020-02-20 21:07:53 +01:00
|
|
|
namespace ROCKSDB_NAMESPACE {
|
2013-12-05 22:09:13 +01:00
|
|
|
|
2015-10-09 01:57:35 +02:00
|
|
|
const uint32_t TablePropertiesCollectorFactory::Context::kUnknownColumnFamily =
|
2022-05-05 22:08:21 +02:00
|
|
|
std::numeric_limits<int32_t>::max();
|
2015-10-09 01:57:35 +02:00
|
|
|
|
2013-12-05 22:09:13 +01:00
|
|
|
namespace {
|
|
|
|
void AppendProperty(
|
|
|
|
std::string& props,
|
|
|
|
const std::string& key,
|
|
|
|
const std::string& value,
|
|
|
|
const std::string& prop_delim,
|
|
|
|
const std::string& kv_delim) {
|
|
|
|
props.append(key);
|
|
|
|
props.append(kv_delim);
|
|
|
|
props.append(value);
|
|
|
|
props.append(prop_delim);
|
|
|
|
}
|
|
|
|
|
|
|
|
template <class TValue>
|
|
|
|
void AppendProperty(
|
|
|
|
std::string& props,
|
|
|
|
const std::string& key,
|
|
|
|
const TValue& value,
|
|
|
|
const std::string& prop_delim,
|
|
|
|
const std::string& kv_delim) {
|
|
|
|
AppendProperty(
|
2014-11-25 05:44:49 +01:00
|
|
|
props, key, ToString(value), prop_delim, kv_delim
|
2013-12-05 22:09:13 +01:00
|
|
|
);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
std::string TableProperties::ToString(
|
|
|
|
const std::string& prop_delim,
|
|
|
|
const std::string& kv_delim) const {
|
|
|
|
std::string result;
|
|
|
|
result.reserve(1024);
|
|
|
|
|
|
|
|
// Basic Info
|
2014-02-05 01:21:47 +01:00
|
|
|
AppendProperty(result, "# data blocks", num_data_blocks, prop_delim,
|
|
|
|
kv_delim);
|
2013-12-05 22:09:13 +01:00
|
|
|
AppendProperty(result, "# entries", num_entries, prop_delim, kv_delim);
|
2018-10-30 23:29:58 +01:00
|
|
|
AppendProperty(result, "# deletions", num_deletions, prop_delim, kv_delim);
|
|
|
|
AppendProperty(result, "# merge operands", num_merge_operands, prop_delim,
|
|
|
|
kv_delim);
|
2018-06-27 05:18:43 +02:00
|
|
|
AppendProperty(result, "# range deletions", num_range_deletions, prop_delim,
|
|
|
|
kv_delim);
|
2013-12-05 22:09:13 +01:00
|
|
|
|
|
|
|
AppendProperty(result, "raw key size", raw_key_size, prop_delim, kv_delim);
|
2014-02-05 01:21:47 +01:00
|
|
|
AppendProperty(result, "raw average key size",
|
|
|
|
num_entries != 0 ? 1.0 * raw_key_size / num_entries : 0.0,
|
|
|
|
prop_delim, kv_delim);
|
|
|
|
AppendProperty(result, "raw value size", raw_value_size, prop_delim,
|
|
|
|
kv_delim);
|
|
|
|
AppendProperty(result, "raw average value size",
|
|
|
|
num_entries != 0 ? 1.0 * raw_value_size / num_entries : 0.0,
|
|
|
|
prop_delim, kv_delim);
|
2013-12-05 22:09:13 +01:00
|
|
|
|
|
|
|
AppendProperty(result, "data block size", data_size, prop_delim, kv_delim);
|
2018-05-26 03:41:31 +02:00
|
|
|
char index_block_size_str[80];
|
|
|
|
snprintf(index_block_size_str, sizeof(index_block_size_str),
|
2018-08-10 01:49:45 +02:00
|
|
|
"index block size (user-key? %d, delta-value? %d)",
|
|
|
|
static_cast<int>(index_key_is_user_key),
|
|
|
|
static_cast<int>(index_value_is_delta_encoded));
|
2018-05-26 03:41:31 +02:00
|
|
|
AppendProperty(result, index_block_size_str, index_size, prop_delim,
|
|
|
|
kv_delim);
|
2017-06-13 19:59:22 +02:00
|
|
|
if (index_partitions != 0) {
|
|
|
|
AppendProperty(result, "# index partitions", index_partitions, prop_delim,
|
|
|
|
kv_delim);
|
|
|
|
AppendProperty(result, "top-level index size", top_level_index_size, prop_delim,
|
|
|
|
kv_delim);
|
|
|
|
}
|
2014-02-05 01:21:47 +01:00
|
|
|
AppendProperty(result, "filter block size", filter_size, prop_delim,
|
|
|
|
kv_delim);
|
2021-05-22 02:10:29 +02:00
|
|
|
AppendProperty(result, "# entries for filter", num_filter_entries, prop_delim,
|
|
|
|
kv_delim);
|
2014-02-05 01:21:47 +01:00
|
|
|
AppendProperty(result, "(estimated) table size",
|
|
|
|
data_size + index_size + filter_size, prop_delim, kv_delim);
|
2013-12-05 22:09:13 +01:00
|
|
|
|
|
|
|
AppendProperty(
|
2014-02-05 01:21:47 +01:00
|
|
|
result, "filter policy name",
|
2013-12-05 22:09:13 +01:00
|
|
|
filter_policy_name.empty() ? std::string("N/A") : filter_policy_name,
|
2014-02-05 01:21:47 +01:00
|
|
|
prop_delim, kv_delim);
|
2013-12-05 22:09:13 +01:00
|
|
|
|
2018-05-21 23:33:55 +02:00
|
|
|
AppendProperty(result, "prefix extractor name",
|
|
|
|
prefix_extractor_name.empty() ? std::string("N/A")
|
|
|
|
: prefix_extractor_name,
|
|
|
|
prop_delim, kv_delim);
|
|
|
|
|
2016-04-09 03:50:18 +02:00
|
|
|
AppendProperty(result, "column family ID",
|
2020-02-20 21:07:53 +01:00
|
|
|
column_family_id ==
|
|
|
|
ROCKSDB_NAMESPACE::TablePropertiesCollectorFactory::
|
|
|
|
Context::kUnknownColumnFamily
|
2016-04-09 03:50:18 +02:00
|
|
|
? std::string("N/A")
|
2020-02-20 21:07:53 +01:00
|
|
|
: ROCKSDB_NAMESPACE::ToString(column_family_id),
|
2016-04-09 03:50:18 +02:00
|
|
|
prop_delim, kv_delim);
|
|
|
|
AppendProperty(
|
|
|
|
result, "column family name",
|
|
|
|
column_family_name.empty() ? std::string("N/A") : column_family_name,
|
|
|
|
prop_delim, kv_delim);
|
|
|
|
|
2016-04-21 19:16:28 +02:00
|
|
|
AppendProperty(result, "comparator name",
|
|
|
|
comparator_name.empty() ? std::string("N/A") : comparator_name,
|
|
|
|
prop_delim, kv_delim);
|
|
|
|
|
|
|
|
AppendProperty(
|
|
|
|
result, "merge operator name",
|
|
|
|
merge_operator_name.empty() ? std::string("N/A") : merge_operator_name,
|
|
|
|
prop_delim, kv_delim);
|
|
|
|
|
|
|
|
AppendProperty(result, "property collectors names",
|
|
|
|
property_collectors_names.empty() ? std::string("N/A")
|
|
|
|
: property_collectors_names,
|
|
|
|
prop_delim, kv_delim);
|
|
|
|
|
2016-05-12 18:47:16 +02:00
|
|
|
AppendProperty(
|
|
|
|
result, "SST file compression algo",
|
|
|
|
compression_name.empty() ? std::string("N/A") : compression_name,
|
|
|
|
prop_delim, kv_delim);
|
|
|
|
|
2019-04-02 23:48:52 +02:00
|
|
|
AppendProperty(
|
|
|
|
result, "SST file compression options",
|
|
|
|
compression_options.empty() ? std::string("N/A") : compression_options,
|
|
|
|
prop_delim, kv_delim);
|
|
|
|
|
2017-06-28 02:02:20 +02:00
|
|
|
AppendProperty(result, "creation time", creation_time, prop_delim, kv_delim);
|
|
|
|
|
2017-10-24 00:22:05 +02:00
|
|
|
AppendProperty(result, "time stamp of earliest key", oldest_key_time,
|
|
|
|
prop_delim, kv_delim);
|
|
|
|
|
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
|
|
|
AppendProperty(result, "file creation time", file_creation_time, prop_delim,
|
|
|
|
kv_delim);
|
|
|
|
|
2021-04-01 03:20:44 +02:00
|
|
|
AppendProperty(result, "slow compression estimated data size",
|
|
|
|
slow_compression_estimated_data_size, prop_delim, kv_delim);
|
|
|
|
AppendProperty(result, "fast compression estimated data size",
|
|
|
|
fast_compression_estimated_data_size, prop_delim, kv_delim);
|
|
|
|
|
2020-06-17 19:55:42 +02:00
|
|
|
// DB identity and DB session ID
|
|
|
|
AppendProperty(result, "DB identity", db_id, prop_delim, kv_delim);
|
|
|
|
AppendProperty(result, "DB session identity", db_session_id, prop_delim,
|
|
|
|
kv_delim);
|
2021-08-21 05:39:52 +02:00
|
|
|
AppendProperty(result, "DB host id", db_host_id, prop_delim, kv_delim);
|
|
|
|
AppendProperty(result, "original file number", orig_file_number, prop_delim,
|
|
|
|
kv_delim);
|
2020-06-17 19:55:42 +02:00
|
|
|
|
Experimental support for SST unique IDs (#8990)
Summary:
* New public header unique_id.h and function GetUniqueIdFromTableProperties
which computes a universally unique identifier based on table properties
of table files from recent RocksDB versions.
* Generation of DB session IDs is refactored so that they are
guaranteed unique in the lifetime of a process running RocksDB.
(SemiStructuredUniqueIdGen, new test included.) Along with file numbers,
this enables SST unique IDs to be guaranteed unique among SSTs generated
in a single process, and "better than random" between processes.
See https://github.com/pdillinger/unique_id
* In addition to public API producing 'external' unique IDs, there is a function
for producing 'internal' unique IDs, with functions for converting between the
two. In short, the external ID is "safe" for things people might do with it, and
the internal ID enables more "power user" features for the future. Specifically,
the external ID goes through a hashing layer so that any subset of bits in the
external ID can be used as a hash of the full ID, while also preserving
uniqueness guarantees in the first 128 bits (bijective both on first 128 bits
and on full 192 bits).
Intended follow-up:
* Use the internal unique IDs in cache keys. (Avoid conflicts with https://github.com/facebook/rocksdb/issues/8912) (The file offset can be XORed into
the third 64-bit value of the unique ID.)
* Publish the external unique IDs in FileStorageInfo (https://github.com/facebook/rocksdb/issues/8968)
Pull Request resolved: https://github.com/facebook/rocksdb/pull/8990
Test Plan:
Unit tests added, and checking of unique ids in stress test.
NOTE in stress test we do not generate nearly enough files to thoroughly
stress uniqueness, but the test trims off pieces of the ID to check for
uniqueness so that we can infer (with some assumptions) stronger
properties in the aggregate.
Reviewed By: zhichao-cao, mrambacher
Differential Revision: D31582865
Pulled By: pdillinger
fbshipit-source-id: 1f620c4c86af9abe2a8d177b9ccf2ad2b9f48243
2021-10-19 08:28:28 +02:00
|
|
|
// Unique ID, when available
|
|
|
|
std::string id;
|
|
|
|
Status s = GetUniqueIdFromTableProperties(*this, &id);
|
|
|
|
AppendProperty(result, "unique ID",
|
|
|
|
s.ok() ? UniqueIdToHumanString(id) : "N/A", prop_delim,
|
|
|
|
kv_delim);
|
|
|
|
|
2013-12-05 22:09:13 +01:00
|
|
|
return result;
|
|
|
|
}
|
|
|
|
|
2015-08-25 21:03:54 +02:00
|
|
|
void TableProperties::Add(const TableProperties& tp) {
|
|
|
|
data_size += tp.data_size;
|
|
|
|
index_size += tp.index_size;
|
2017-06-13 19:59:22 +02:00
|
|
|
index_partitions += tp.index_partitions;
|
|
|
|
top_level_index_size += tp.top_level_index_size;
|
2018-05-26 03:41:31 +02:00
|
|
|
index_key_is_user_key += tp.index_key_is_user_key;
|
2018-08-10 01:49:45 +02:00
|
|
|
index_value_is_delta_encoded += tp.index_value_is_delta_encoded;
|
2015-08-25 21:03:54 +02:00
|
|
|
filter_size += tp.filter_size;
|
|
|
|
raw_key_size += tp.raw_key_size;
|
|
|
|
raw_value_size += tp.raw_value_size;
|
|
|
|
num_data_blocks += tp.num_data_blocks;
|
|
|
|
num_entries += tp.num_entries;
|
2021-05-22 02:10:29 +02:00
|
|
|
num_filter_entries += tp.num_filter_entries;
|
2018-10-30 23:29:58 +01:00
|
|
|
num_deletions += tp.num_deletions;
|
|
|
|
num_merge_operands += tp.num_merge_operands;
|
2018-06-27 05:18:43 +02:00
|
|
|
num_range_deletions += tp.num_range_deletions;
|
2021-04-01 03:20:44 +02:00
|
|
|
slow_compression_estimated_data_size +=
|
|
|
|
tp.slow_compression_estimated_data_size;
|
|
|
|
fast_compression_estimated_data_size +=
|
|
|
|
tp.fast_compression_estimated_data_size;
|
2015-08-25 21:03:54 +02:00
|
|
|
}
|
|
|
|
|
2020-12-19 16:59:08 +01:00
|
|
|
std::map<std::string, uint64_t>
|
|
|
|
TableProperties::GetAggregatablePropertiesAsMap() const {
|
|
|
|
std::map<std::string, uint64_t> rv;
|
|
|
|
rv["data_size"] = data_size;
|
|
|
|
rv["index_size"] = index_size;
|
|
|
|
rv["index_partitions"] = index_partitions;
|
|
|
|
rv["top_level_index_size"] = top_level_index_size;
|
|
|
|
rv["filter_size"] = filter_size;
|
|
|
|
rv["raw_key_size"] = raw_key_size;
|
|
|
|
rv["raw_value_size"] = raw_value_size;
|
|
|
|
rv["num_data_blocks"] = num_data_blocks;
|
|
|
|
rv["num_entries"] = num_entries;
|
2021-05-22 02:10:29 +02:00
|
|
|
rv["num_filter_entries"] = num_filter_entries;
|
2020-12-19 16:59:08 +01:00
|
|
|
rv["num_deletions"] = num_deletions;
|
|
|
|
rv["num_merge_operands"] = num_merge_operands;
|
|
|
|
rv["num_range_deletions"] = num_range_deletions;
|
2021-04-01 03:20:44 +02:00
|
|
|
rv["slow_compression_estimated_data_size"] =
|
|
|
|
slow_compression_estimated_data_size;
|
|
|
|
rv["fast_compression_estimated_data_size"] =
|
|
|
|
fast_compression_estimated_data_size;
|
2020-12-19 16:59:08 +01:00
|
|
|
return rv;
|
|
|
|
}
|
|
|
|
|
2022-04-06 19:33:00 +02:00
|
|
|
// WARNING: manual update to this function is needed
|
|
|
|
// whenever a new string property is added to TableProperties
|
|
|
|
// to reduce approximation error.
|
|
|
|
//
|
|
|
|
// TODO: eliminate the need of manually updating this function
|
|
|
|
// for new string properties
|
|
|
|
std::size_t TableProperties::ApproximateMemoryUsage() const {
|
|
|
|
std::size_t usage = 0;
|
|
|
|
#ifdef ROCKSDB_MALLOC_USABLE_SIZE
|
|
|
|
usage += malloc_usable_size((void*)this);
|
|
|
|
#else
|
|
|
|
usage += sizeof(*this);
|
|
|
|
#endif // ROCKSDB_MALLOC_USABLE_SIZE
|
|
|
|
|
|
|
|
std::size_t string_props_mem_usage =
|
|
|
|
db_id.size() + db_session_id.size() + db_host_id.size() +
|
|
|
|
column_family_name.size() + filter_policy_name.size() +
|
|
|
|
comparator_name.size() + merge_operator_name.size() +
|
|
|
|
prefix_extractor_name.size() + property_collectors_names.size() +
|
|
|
|
compression_name.size() + compression_options.size();
|
|
|
|
usage += string_props_mem_usage;
|
|
|
|
|
|
|
|
for (auto iter = user_collected_properties.begin();
|
|
|
|
iter != user_collected_properties.end(); ++iter) {
|
|
|
|
usage += (iter->first.size() + iter->second.size());
|
|
|
|
}
|
|
|
|
|
|
|
|
return usage;
|
|
|
|
}
|
|
|
|
|
2020-06-17 19:55:42 +02:00
|
|
|
const std::string TablePropertiesNames::kDbId = "rocksdb.creating.db.identity";
|
|
|
|
const std::string TablePropertiesNames::kDbSessionId =
|
|
|
|
"rocksdb.creating.session.identity";
|
2020-10-19 20:37:05 +02:00
|
|
|
const std::string TablePropertiesNames::kDbHostId =
|
|
|
|
"rocksdb.creating.host.identity";
|
2021-08-21 05:39:52 +02:00
|
|
|
const std::string TablePropertiesNames::kOriginalFileNumber =
|
|
|
|
"rocksdb.original.file.number";
|
2013-12-05 22:09:13 +01:00
|
|
|
const std::string TablePropertiesNames::kDataSize =
|
|
|
|
"rocksdb.data.size";
|
|
|
|
const std::string TablePropertiesNames::kIndexSize =
|
|
|
|
"rocksdb.index.size";
|
2017-06-13 19:59:22 +02:00
|
|
|
const std::string TablePropertiesNames::kIndexPartitions =
|
|
|
|
"rocksdb.index.partitions";
|
|
|
|
const std::string TablePropertiesNames::kTopLevelIndexSize =
|
|
|
|
"rocksdb.top-level.index.size";
|
2018-05-26 03:41:31 +02:00
|
|
|
const std::string TablePropertiesNames::kIndexKeyIsUserKey =
|
|
|
|
"rocksdb.index.key.is.user.key";
|
2018-08-10 01:49:45 +02:00
|
|
|
const std::string TablePropertiesNames::kIndexValueIsDeltaEncoded =
|
|
|
|
"rocksdb.index.value.is.delta.encoded";
|
2013-12-05 22:09:13 +01:00
|
|
|
const std::string TablePropertiesNames::kFilterSize =
|
|
|
|
"rocksdb.filter.size";
|
|
|
|
const std::string TablePropertiesNames::kRawKeySize =
|
|
|
|
"rocksdb.raw.key.size";
|
|
|
|
const std::string TablePropertiesNames::kRawValueSize =
|
|
|
|
"rocksdb.raw.value.size";
|
|
|
|
const std::string TablePropertiesNames::kNumDataBlocks =
|
|
|
|
"rocksdb.num.data.blocks";
|
|
|
|
const std::string TablePropertiesNames::kNumEntries =
|
|
|
|
"rocksdb.num.entries";
|
2021-05-22 02:10:29 +02:00
|
|
|
const std::string TablePropertiesNames::kNumFilterEntries =
|
|
|
|
"rocksdb.num.filter_entries";
|
2018-10-30 23:29:58 +01:00
|
|
|
const std::string TablePropertiesNames::kDeletedKeys = "rocksdb.deleted.keys";
|
|
|
|
const std::string TablePropertiesNames::kMergeOperands =
|
|
|
|
"rocksdb.merge.operands";
|
2018-06-27 05:18:43 +02:00
|
|
|
const std::string TablePropertiesNames::kNumRangeDeletions =
|
|
|
|
"rocksdb.num.range-deletions";
|
2013-12-05 22:09:13 +01:00
|
|
|
const std::string TablePropertiesNames::kFilterPolicy =
|
|
|
|
"rocksdb.filter.policy";
|
2013-12-20 18:35:24 +01:00
|
|
|
const std::string TablePropertiesNames::kFormatVersion =
|
|
|
|
"rocksdb.format.version";
|
|
|
|
const std::string TablePropertiesNames::kFixedKeyLen =
|
|
|
|
"rocksdb.fixed.key.length";
|
2016-04-07 08:10:32 +02:00
|
|
|
const std::string TablePropertiesNames::kColumnFamilyId =
|
|
|
|
"rocksdb.column.family.id";
|
|
|
|
const std::string TablePropertiesNames::kColumnFamilyName =
|
|
|
|
"rocksdb.column.family.name";
|
2016-04-21 19:16:28 +02:00
|
|
|
const std::string TablePropertiesNames::kComparator = "rocksdb.comparator";
|
|
|
|
const std::string TablePropertiesNames::kMergeOperator =
|
|
|
|
"rocksdb.merge.operator";
|
2016-08-26 20:46:32 +02:00
|
|
|
const std::string TablePropertiesNames::kPrefixExtractorName =
|
|
|
|
"rocksdb.prefix.extractor.name";
|
2016-04-21 19:16:28 +02:00
|
|
|
const std::string TablePropertiesNames::kPropertyCollectors =
|
|
|
|
"rocksdb.property.collectors";
|
2016-05-12 18:47:16 +02:00
|
|
|
const std::string TablePropertiesNames::kCompression = "rocksdb.compression";
|
2019-04-02 23:48:52 +02:00
|
|
|
const std::string TablePropertiesNames::kCompressionOptions =
|
|
|
|
"rocksdb.compression_options";
|
2017-06-28 02:02:20 +02:00
|
|
|
const std::string TablePropertiesNames::kCreationTime = "rocksdb.creation.time";
|
2017-10-24 00:22:05 +02:00
|
|
|
const std::string TablePropertiesNames::kOldestKeyTime =
|
|
|
|
"rocksdb.oldest.key.time";
|
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
|
|
|
const std::string TablePropertiesNames::kFileCreationTime =
|
|
|
|
"rocksdb.file.creation.time";
|
2021-04-01 03:20:44 +02:00
|
|
|
const std::string TablePropertiesNames::kSlowCompressionEstimatedDataSize =
|
|
|
|
"rocksdb.sample_for_compression.slow.data.size";
|
|
|
|
const std::string TablePropertiesNames::kFastCompressionEstimatedDataSize =
|
|
|
|
"rocksdb.sample_for_compression.fast.data.size";
|
2013-12-05 22:09:13 +01:00
|
|
|
|
Experimental support for SST unique IDs (#8990)
Summary:
* New public header unique_id.h and function GetUniqueIdFromTableProperties
which computes a universally unique identifier based on table properties
of table files from recent RocksDB versions.
* Generation of DB session IDs is refactored so that they are
guaranteed unique in the lifetime of a process running RocksDB.
(SemiStructuredUniqueIdGen, new test included.) Along with file numbers,
this enables SST unique IDs to be guaranteed unique among SSTs generated
in a single process, and "better than random" between processes.
See https://github.com/pdillinger/unique_id
* In addition to public API producing 'external' unique IDs, there is a function
for producing 'internal' unique IDs, with functions for converting between the
two. In short, the external ID is "safe" for things people might do with it, and
the internal ID enables more "power user" features for the future. Specifically,
the external ID goes through a hashing layer so that any subset of bits in the
external ID can be used as a hash of the full ID, while also preserving
uniqueness guarantees in the first 128 bits (bijective both on first 128 bits
and on full 192 bits).
Intended follow-up:
* Use the internal unique IDs in cache keys. (Avoid conflicts with https://github.com/facebook/rocksdb/issues/8912) (The file offset can be XORed into
the third 64-bit value of the unique ID.)
* Publish the external unique IDs in FileStorageInfo (https://github.com/facebook/rocksdb/issues/8968)
Pull Request resolved: https://github.com/facebook/rocksdb/pull/8990
Test Plan:
Unit tests added, and checking of unique ids in stress test.
NOTE in stress test we do not generate nearly enough files to thoroughly
stress uniqueness, but the test trims off pieces of the ID to check for
uniqueness so that we can infer (with some assumptions) stronger
properties in the aggregate.
Reviewed By: zhichao-cao, mrambacher
Differential Revision: D31582865
Pulled By: pdillinger
fbshipit-source-id: 1f620c4c86af9abe2a8d177b9ccf2ad2b9f48243
2021-10-19 08:28:28 +02:00
|
|
|
#ifndef NDEBUG
|
2022-04-07 21:25:43 +02:00
|
|
|
// WARNING: TEST_SetRandomTableProperties assumes the following layout of
|
|
|
|
// TableProperties
|
|
|
|
//
|
|
|
|
// struct TableProperties {
|
|
|
|
// int64_t orig_file_number = 0;
|
|
|
|
// ...
|
|
|
|
// ... int64_t properties only
|
|
|
|
// ...
|
|
|
|
// std::string db_id;
|
|
|
|
// ...
|
|
|
|
// ... std::string properties only
|
|
|
|
// ...
|
|
|
|
// std::string compression_options;
|
|
|
|
// UserCollectedProperties user_collected_properties;
|
|
|
|
// ...
|
|
|
|
// ... Other extra properties: non-int64_t/non-std::string properties only
|
|
|
|
// ...
|
|
|
|
// }
|
Experimental support for SST unique IDs (#8990)
Summary:
* New public header unique_id.h and function GetUniqueIdFromTableProperties
which computes a universally unique identifier based on table properties
of table files from recent RocksDB versions.
* Generation of DB session IDs is refactored so that they are
guaranteed unique in the lifetime of a process running RocksDB.
(SemiStructuredUniqueIdGen, new test included.) Along with file numbers,
this enables SST unique IDs to be guaranteed unique among SSTs generated
in a single process, and "better than random" between processes.
See https://github.com/pdillinger/unique_id
* In addition to public API producing 'external' unique IDs, there is a function
for producing 'internal' unique IDs, with functions for converting between the
two. In short, the external ID is "safe" for things people might do with it, and
the internal ID enables more "power user" features for the future. Specifically,
the external ID goes through a hashing layer so that any subset of bits in the
external ID can be used as a hash of the full ID, while also preserving
uniqueness guarantees in the first 128 bits (bijective both on first 128 bits
and on full 192 bits).
Intended follow-up:
* Use the internal unique IDs in cache keys. (Avoid conflicts with https://github.com/facebook/rocksdb/issues/8912) (The file offset can be XORed into
the third 64-bit value of the unique ID.)
* Publish the external unique IDs in FileStorageInfo (https://github.com/facebook/rocksdb/issues/8968)
Pull Request resolved: https://github.com/facebook/rocksdb/pull/8990
Test Plan:
Unit tests added, and checking of unique ids in stress test.
NOTE in stress test we do not generate nearly enough files to thoroughly
stress uniqueness, but the test trims off pieces of the ID to check for
uniqueness so that we can infer (with some assumptions) stronger
properties in the aggregate.
Reviewed By: zhichao-cao, mrambacher
Differential Revision: D31582865
Pulled By: pdillinger
fbshipit-source-id: 1f620c4c86af9abe2a8d177b9ccf2ad2b9f48243
2021-10-19 08:28:28 +02:00
|
|
|
void TEST_SetRandomTableProperties(TableProperties* props) {
|
|
|
|
Random* r = Random::GetTLSInstance();
|
|
|
|
uint64_t* pu = &props->orig_file_number;
|
|
|
|
assert(static_cast<void*>(pu) == static_cast<void*>(props));
|
|
|
|
std::string* ps = &props->db_id;
|
|
|
|
const uint64_t* const pu_end = reinterpret_cast<const uint64_t*>(ps);
|
2022-04-07 21:25:43 +02:00
|
|
|
// Use the last string property's address instead of
|
|
|
|
// the first extra property (e.g `user_collected_properties`)'s address
|
|
|
|
// in the for-loop to avoid advancing pointer to pointing to
|
|
|
|
// potential non-zero padding bytes between these two addresses due to
|
|
|
|
// user_collected_properties's alignment requirement
|
|
|
|
const std::string* const ps_end_inclusive = &props->compression_options;
|
Experimental support for SST unique IDs (#8990)
Summary:
* New public header unique_id.h and function GetUniqueIdFromTableProperties
which computes a universally unique identifier based on table properties
of table files from recent RocksDB versions.
* Generation of DB session IDs is refactored so that they are
guaranteed unique in the lifetime of a process running RocksDB.
(SemiStructuredUniqueIdGen, new test included.) Along with file numbers,
this enables SST unique IDs to be guaranteed unique among SSTs generated
in a single process, and "better than random" between processes.
See https://github.com/pdillinger/unique_id
* In addition to public API producing 'external' unique IDs, there is a function
for producing 'internal' unique IDs, with functions for converting between the
two. In short, the external ID is "safe" for things people might do with it, and
the internal ID enables more "power user" features for the future. Specifically,
the external ID goes through a hashing layer so that any subset of bits in the
external ID can be used as a hash of the full ID, while also preserving
uniqueness guarantees in the first 128 bits (bijective both on first 128 bits
and on full 192 bits).
Intended follow-up:
* Use the internal unique IDs in cache keys. (Avoid conflicts with https://github.com/facebook/rocksdb/issues/8912) (The file offset can be XORed into
the third 64-bit value of the unique ID.)
* Publish the external unique IDs in FileStorageInfo (https://github.com/facebook/rocksdb/issues/8968)
Pull Request resolved: https://github.com/facebook/rocksdb/pull/8990
Test Plan:
Unit tests added, and checking of unique ids in stress test.
NOTE in stress test we do not generate nearly enough files to thoroughly
stress uniqueness, but the test trims off pieces of the ID to check for
uniqueness so that we can infer (with some assumptions) stronger
properties in the aggregate.
Reviewed By: zhichao-cao, mrambacher
Differential Revision: D31582865
Pulled By: pdillinger
fbshipit-source-id: 1f620c4c86af9abe2a8d177b9ccf2ad2b9f48243
2021-10-19 08:28:28 +02:00
|
|
|
|
|
|
|
for (; pu < pu_end; ++pu) {
|
|
|
|
*pu = r->Next64();
|
|
|
|
}
|
|
|
|
assert(static_cast<void*>(pu) == static_cast<void*>(ps));
|
2022-04-07 21:25:43 +02:00
|
|
|
for (; ps <= ps_end_inclusive; ++ps) {
|
Experimental support for SST unique IDs (#8990)
Summary:
* New public header unique_id.h and function GetUniqueIdFromTableProperties
which computes a universally unique identifier based on table properties
of table files from recent RocksDB versions.
* Generation of DB session IDs is refactored so that they are
guaranteed unique in the lifetime of a process running RocksDB.
(SemiStructuredUniqueIdGen, new test included.) Along with file numbers,
this enables SST unique IDs to be guaranteed unique among SSTs generated
in a single process, and "better than random" between processes.
See https://github.com/pdillinger/unique_id
* In addition to public API producing 'external' unique IDs, there is a function
for producing 'internal' unique IDs, with functions for converting between the
two. In short, the external ID is "safe" for things people might do with it, and
the internal ID enables more "power user" features for the future. Specifically,
the external ID goes through a hashing layer so that any subset of bits in the
external ID can be used as a hash of the full ID, while also preserving
uniqueness guarantees in the first 128 bits (bijective both on first 128 bits
and on full 192 bits).
Intended follow-up:
* Use the internal unique IDs in cache keys. (Avoid conflicts with https://github.com/facebook/rocksdb/issues/8912) (The file offset can be XORed into
the third 64-bit value of the unique ID.)
* Publish the external unique IDs in FileStorageInfo (https://github.com/facebook/rocksdb/issues/8968)
Pull Request resolved: https://github.com/facebook/rocksdb/pull/8990
Test Plan:
Unit tests added, and checking of unique ids in stress test.
NOTE in stress test we do not generate nearly enough files to thoroughly
stress uniqueness, but the test trims off pieces of the ID to check for
uniqueness so that we can infer (with some assumptions) stronger
properties in the aggregate.
Reviewed By: zhichao-cao, mrambacher
Differential Revision: D31582865
Pulled By: pdillinger
fbshipit-source-id: 1f620c4c86af9abe2a8d177b9ccf2ad2b9f48243
2021-10-19 08:28:28 +02:00
|
|
|
*ps = r->RandomBinaryString(13);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
|
2020-02-20 21:07:53 +01:00
|
|
|
} // namespace ROCKSDB_NAMESPACE
|