mirrors/scylladb
1
0
Fork 0
mirror of https://github.com/scylladb/scylladb.git synced 2026-04-26 19:35:12 +00:00
Code Issues Packages Projects Releases Wiki Activity
Files
7cdda510eef4e2fd5374c087b96ffaeb3948cd98
scylladb/sstables
History
Lakshmi Narayanan Sreethar 1d1e572962 sstables: skip bloom filter rebuilds with minimal savings 
If a bloom filter was built with a bad partition estimate, it is rebuilt
right before the sstable is sealed. The rebuild is already skipped if
the current bitset size results in a false-positive rate within 75%–125%
of the configured value.

This patch adds additional conditions to prevent rebuilds when the
savings are minimal. It also skips rebuilding for garbage collected
sstables, since they will be dropped soon anyway.

Also updated and added more test cases to cover these new criteria for
bloom filter rebuilds.

Fixes #25464
Fixes #25468

Signed-off-by: Lakshmi Narayanan Sreethar <lakshmi.sreethar@scylladb.com>

Closes scylladb/scylladb#25968
2025-09-14 18:19:50 +03:00
..
kl
Merge 'sstables/mx/reader: accommodate inexact partition indexes' from Michał Chojnowski
2025-07-27 19:39:36 +03:00
mx
sstables/writer: add an accessor for the current write position in Data.db
2025-09-07 00:32:02 +02:00
trie
treewide: fix spelling errors
2025-09-12 15:58:19 +03:00
abstract_index_reader.hh
sstables/trie: introduce bti_index_reader
2025-09-07 00:32:02 +02:00
basic_info.hh
system_keyspace: Upgrade compaction_history table
2025-05-14 08:32:05 +02:00
binary_search.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
checksum_utils.hh
utils: do not include unused headers
2025-01-14 07:56:39 -05:00
checksummed_data_source.cc
sstables: Start using make_data_or_index_source in sstable
2025-07-15 10:10:23 +03:00
checksummed_data_source.hh
sstables: Start using make_data_or_index_source in sstable
2025-07-15 10:10:23 +03:00
CMakeLists.txt
sstables/trie: introduce bti_index_reader
2025-09-07 00:32:02 +02:00
column_translation.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
component_type.hh
db: snapshot: backup_task: do_backup: organize components by sstable generation
2025-04-09 08:54:06 +03:00
compress.cc
compress: move compress.cc/hh to sstables/compressor
2025-07-31 13:10:41 +03:00
compress.hh
compress: move compress.cc/hh to sstables/compressor
2025-07-31 13:10:41 +03:00
compressor.cc
compress: move compress.cc/hh to sstables/compressor
2025-07-31 13:10:41 +03:00
compressor.hh
compress: move compress.cc/hh to sstables/compressor
2025-07-31 13:10:41 +03:00
consumer.hh
sstables/consumer: add read_56()
2025-09-07 00:30:15 +02:00
data_source_types.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
disk_types.hh
sstables: disk_types: disk_set_of_tagged_union: boost::variant -> std::variant
2025-01-27 09:29:26 -05:00
downsampling.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
exceptions.hh
sstables/trie: implement BTI node traversal
2025-08-05 00:56:50 +02:00
file_writer.hh
sstables: Make file_writer keep component_name on board
2025-03-19 13:03:29 +03:00
filter.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
generation_type.hh
sstables: sstable_generation_generator: stop tracking highest generation
2025-08-08 11:46:21 +03:00
hyperloglog.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
index_entry.hh
sstables: add sstable::make_index_reader() and use where appropriate
2025-07-17 10:32:57 +02:00
index_reader.hh
sstables: extract abstract_index_reader from index_reader.hh to its own header
2025-09-07 00:30:15 +02:00
integrity_checked_file_impl.cc
sstables: switch from boost to std::ranges::all_of()
2025-02-10 15:44:55 +03:00
integrity_checked_file_impl.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
key.hh
treewide: Move keys related files to a new keys directory
2025-07-25 10:45:32 +03:00
liveness_info.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
m_format_read_helpers.cc
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
m_format_read_helpers.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
metadata_collector.cc
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
metadata_collector.hh
sstables: Add set_repaired_at to metadata_collector
2025-08-11 10:10:08 +08:00
mutation_fragment_filter.hh
treewide: Move keys related files to a new keys directory
2025-07-25 10:45:32 +03:00
open_info.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
partition_index_cache_stats.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
partition_index_cache.hh
sstables: replace SCYLLA_ASSERT() with parse_assert() on the read path
2025-06-24 09:16:28 +03:00
prepended_input_stream.cc
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
prepended_input_stream.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
processing_result_generator.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
progress_monitor.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
promoted_index_blocks_reader.hh
sstables: replace SCYLLA_ASSERT() with parse_assert() on the read path
2025-06-24 09:16:28 +03:00
random_access_reader.cc
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
random_access_reader.hh
sstables: replace SCYLLA_ASSERT() with parse_assert() on the read path
2025-06-24 09:16:28 +03:00
scanning_clustered_index_cursor.hh
sstables: replace SCYLLA_ASSERT() with parse_assert() on the read path
2025-06-24 09:16:28 +03:00
segmented_compress_params.hh
treewide: Add missing #pragma once
2025-09-01 14:58:21 +03:00
shareable_components.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
shared_sstable.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
sstable_compressor_factory.hh
compress: add some test-only APIs
2025-05-07 14:43:20 +02:00
sstable_directory.cc
sstables: sstable_directory: stop tracking highest_generation
2025-08-08 11:46:21 +03:00
sstable_directory.hh
sstables: sstable_directory: stop tracking highest_generation
2025-08-08 11:46:21 +03:00
sstable_mutation_reader.cc
db,sstables: migate boost::range::stable_partition to std library
2024-12-19 14:56:07 +02:00
sstable_mutation_reader.hh
Merge 'sstables/mx/reader: accommodate inexact partition indexes' from Michał Chojnowski
2025-07-27 19:39:36 +03:00
sstable_set_impl.hh
sstables: Fix quadratic space complexity in partitioned_sstable_set
2025-04-29 15:47:33 -03:00
sstable_set.cc
treewide: Rename table_state to compaction_group_view
2025-08-08 06:51:28 +03:00
sstable_set.hh
sstables: Fix quadratic space complexity in partitioned_sstable_set
2025-04-29 15:47:33 -03:00
sstable_version_k_l.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
sstable_version_m.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
sstable_version.cc
sstables: Move versions static-assertion check to .cc file
2025-07-07 13:16:21 +03:00
sstable_version.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
sstable_writer.hh
sstables/writer: add an accessor for the current write position in Data.db
2025-09-07 00:32:02 +02:00
sstables_manager_subscription.hh
sstables_manager: add subscriptions
2025-04-09 08:54:07 +03:00
sstables_manager.cc
s3: Export memory usage gauge (metrics)
2025-09-05 18:25:34 +03:00
sstables_manager.hh
s3: Export memory usage gauge (metrics)
2025-09-05 18:25:34 +03:00
sstables_registry.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
sstables.cc
sstables: skip bloom filter rebuilds with minimal savings
2025-09-14 18:19:50 +03:00
sstables.hh
compaction: Add tablet incremental repair support
2025-08-18 11:01:21 +08:00
stats.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
storage.cc
sstables: add make_data_or_index_source to the storage
2025-07-06 09:18:39 +03:00
storage.hh
sstables: add make_data_or_index_source to the storage
2025-07-06 09:18:39 +03:00
types_fwd.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
types.hh
sstables/types: make deletion_time::operator tombstone() const
2025-09-07 00:30:15 +02:00
version.hh
sstables: version: drop custom operator<=>
2025-07-22 12:39:01 +02:00
writer_impl.hh
sstables/writer: add an accessor for the current write position in Data.db
2025-09-07 00:32:02 +02:00
writer.cc
sstables/writer: add an accessor for the current write position in Data.db
2025-09-07 00:32:02 +02:00
writer.hh
compress: remove compressor::option_names()
2025-04-01 00:07:27 +02:00