mirrors/scylladb
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
Files
2d2a2ef277725b06ca2dd10ff6abb87be2edfdbf
scylladb/utils
History
Pavel Emelyanov b56d6fbb84 Merge 'sstables: Fix quadratic space complexity in partitioned_sstable_set' from Raphael Raph Carvalho 
Interval map is very susceptible to quadratic space behavior when it's flooded with many entries overlapping all (or most of) intervals, since each such entry will have presence on all intervals it overlaps with.

A trigger we observed was memtable flush storm, which creates many small "L0" sstables that spans roughly the entire token range.

Since we cannot rely on insertion order, solution will be about storing sstables with such wide ranges in a vector (unleveled).

There should be no consequence for single-key reads, since upper layer applies an additional filtering based on token of key being queried.
And for range scans, there can be an increase in memory usage, but not significant because the sstables span an wide range and would have been selected in the combined reader if the range of scan overlaps with them.

Anyway, this is a protection against storm of memtable flushes and shouldn't be the common scenario.

It works both with tablets and vnodes, by adjusting the token range spanned by compaction group accordingly.

Fixes #23634.

We can backport this into 2024.2, 2025.1, but we should let this cook in master for 1 month or so.

Closes scylladb/scylladb#23806

* github.com:scylladb/scylladb:
  test: Verify partitioned set store split and unsplit correctly
  sstables: Fix quadratic space complexity in partitioned_sstable_set
  compaction: Wire table_state into make_sstable_set()
  compaction: Introduce token_range() to table_state
  dht: Add overlap_ratio() for token range
2025-05-05 11:28:38 +03:00
..
abi
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
arch/powerpc/crc32-vpmsum
…
gz
utils: do not include unused headers
2025-01-14 07:56:39 -05:00
lsa
utils: do not include unused headers
2025-01-14 07:56:39 -05:00
s3
Merge 'Add CopyObject API implementation to S3 client' from Ernest Zaslavsky
2025-04-18 16:17:41 +03:00
advanced_rpc_compressor_protocol.hh
utils: add advanced_rpc_compressor
2024-12-23 23:37:02 +01:00
advanced_rpc_compressor.cc
treewide: use angle brackets for including seastar headers
2025-03-17 10:03:06 +02:00
advanced_rpc_compressor.hh
utils: use std::to_underlying() when appropriate
2025-02-19 12:12:28 +03:00
alien_worker.cc
utils: introduce alien_worker
2024-12-23 23:37:02 +01:00
alien_worker.hh
utils: introduce alien_worker
2024-12-23 23:37:02 +01:00
allocation_strategy.hh
utils: do not include unused headers
2025-01-14 07:56:39 -05:00
amortized_reserve.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
anchorless_list.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
array-search.cc
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
array-search.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
ascii.cc
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
ascii.hh
utils: do not include unused headers
2025-01-14 07:56:39 -05:00
assert.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
atomic_vector.hh
utils: do not include unused headers
2025-01-14 07:56:39 -05:00
aws_sigv4.cc
utils: do not include unused headers
2025-01-14 07:56:39 -05:00
aws_sigv4.hh
utils: do not include unused headers
2025-01-14 07:56:39 -05:00
base64.cc
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
base64.hh
utils: do not include unused headers
2025-01-14 07:56:39 -05:00
big_decimal.cc
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
big_decimal.hh
utils: do not include unused headers
2025-01-14 07:56:39 -05:00
bit_cast.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
bloom_calculations.cc
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
bloom_calculations.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
bloom_filter.cc
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
bloom_filter.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
bounded_stats_deque.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
bptree.hh
B+tree: Clean const_iterator->iterator conversion
2025-02-26 23:17:27 +02:00
buffer_input_stream.cc
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
buffer_input_stream.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
buffer_view-to-managed_bytes_view.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
build_id.cc
treewide: change assert() to SCYLLA_ASSERT()
2024-08-05 08:23:35 +03:00
build_id.hh
…
cached_file_stats.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
cached_file.hh
utils: do not include unused headers
2025-01-14 07:56:39 -05:00
checked-file-impl.hh
code: Move checked-file-impl.hh to util/
2025-03-06 10:22:05 +02:00
chunked_vector.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
class_registrator.hh
utils: do not include unused headers
2025-01-14 07:56:39 -05:00
clmul.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
CMakeLists.txt
Merge 'Fix EAR not applied on write to S3 (but on read).' from Calle Wilund
2025-03-24 10:12:46 +03:00
coarse_steady_clock.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
collection-concepts.hh
utils: do not include unused headers
2025-01-14 07:56:39 -05:00
compact-radix-tree.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
composite_abort_source.hh
treewide: change assert() to SCYLLA_ASSERT()
2024-08-05 08:23:35 +03:00
config_file_impl.hh
tree: Remove unused boost headers
2025-02-25 10:32:32 +03:00
config_file.cc
tree: Remove unused boost headers
2025-02-25 10:32:32 +03:00
config_file.hh
Merge 'config: prevent SIGHUP from changing non-liveupdatable parameters' from Andrzej Jackowski
2025-02-06 11:33:59 +03:00
contiguous_shared_buffer.hh
utils: do not include unused headers
2025-01-14 07:56:39 -05:00
coroutine.hh
utils: do not include unused headers
2025-01-14 07:56:39 -05:00
crc.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
cross-shard-barrier.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
data_input.hh
utils: do not include unused headers
2025-01-14 07:56:39 -05:00
date.h
date: do not put space before literal operator
2024-10-21 11:21:52 +03:00
dict_trainer.cc
utils/hashers: add get_sha256()
2025-04-01 00:07:28 +02:00
dict_trainer.hh
utils: do not include unused headers
2025-01-14 07:56:39 -05:00
digest_algorithm.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
digester.hh
utils: do not include unused headers
2025-01-14 07:56:39 -05:00
directories.cc
tree: Fix polymorphic exception handling by using references
2025-02-26 23:15:16 +02:00
directories.hh
utils: do not include unused headers
2025-01-14 07:56:39 -05:00
disk_space_monitor.cc
utils: phased_barrier, pluggable: use named gate
2025-04-12 11:47:00 +03:00
disk_space_monitor.hh
config, disk_space_monitor: Allow overriding capacity via config
2025-03-06 13:35:37 +01:00
disk-error-handler.cc
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
disk-error-handler.hh
utils: do not include unused headers
2025-01-14 07:56:39 -05:00
div_ceil.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
double-decker.hh
B+tree: Clean const_iterator->iterator conversion
2025-02-26 23:17:27 +02:00
dynamic_bitset.cc
treewide: migrate from boost::adaptors::reversed to std::views::reverse
2025-01-07 13:22:00 +02:00
dynamic_bitset.hh
utils: do not include unused headers
2025-01-14 07:56:39 -05:00
entangled.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
enum_option.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
error_injection.cc
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
error_injection.hh
test: Verify partitioned set store split and unsplit correctly
2025-04-29 15:47:33 -03:00
estimated_histogram.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
exception_container.hh
utils: do not include unused headers
2025-01-14 07:56:39 -05:00
exceptions.cc
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
exceptions.hh
exceptions: Add try_catch_nested to universally handle nested exceptions of the same type.
2025-03-26 11:15:13 +01:00
exponential_backoff_retry.hh
utils: do not include unused headers
2025-01-14 07:56:39 -05:00
extremum_tracking.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
file_lock.cc
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
file_lock.hh
utils: do not include unused headers
2025-01-14 07:56:39 -05:00
flush_queue.hh
utils: flush_queue: use named_gate
2025-04-12 11:47:02 +03:00
fragment_range.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
fragmented_temporary_buffer.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
from_chars_exactly.hh
utils: introduce from_chars_exactly()
2025-01-09 11:14:49 +02:00
hash.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
hashers.cc
utils/hashers: add get_sha256()
2025-04-01 00:07:28 +02:00
hashers.hh
utils/hashers: add get_sha256()
2025-04-01 00:07:28 +02:00
hashing.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
histogram_metrics_helper.cc
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
histogram_metrics_helper.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
histogram.hh
utils: do not include unused headers
2025-01-14 07:56:39 -05:00
http.hh
utils: do not include unused headers
2025-01-14 07:56:39 -05:00
human_readable.cc
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
human_readable.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
i_filter.cc
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
i_filter.hh
utils: do not include unused headers
2025-01-14 07:56:39 -05:00
immutable-collection.hh
utils: do not include unused headers
2025-01-14 07:56:39 -05:00
input_stream.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
int_range.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
intrusive_btree.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
intrusive-array.hh
utils: do not include unused headers
2025-01-14 07:56:39 -05:00
io-wrappers.cc
utils: Add "io-wrappers", useful IO helper types
2025-03-20 14:54:09 +00:00
io-wrappers.hh
utils: Add "io-wrappers", useful IO helper types
2025-03-20 14:54:09 +00:00
labels.cc
Adding the __level and features labels
2025-03-03 16:58:38 +02:00
labels.hh
Adding the __level and features labels
2025-03-03 16:58:38 +02:00
large_bitset.cc
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
large_bitset.hh
utils: do not include unused headers
2025-01-14 07:56:39 -05:00
latency.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
lexicographical_compare.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
like_matcher.cc
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
like_matcher.hh
utils: do not include unused headers
2025-01-14 07:56:39 -05:00
limiting_data_source.cc
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
limiting_data_source.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
linearizing_input_stream.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
lister.cc
code: Move checked-file-impl.hh to util/
2025-03-06 10:22:05 +02:00
lister.hh
utils: do not include unused headers
2025-01-14 07:56:39 -05:00
loading_cache.hh
utils::loading_cache: gracefully skip timer if gate closed
2025-04-30 16:43:22 +03:00
loading_shared_values.hh
utils: do not include unused headers
2025-01-14 07:56:39 -05:00
log_heap.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
log.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
logalloc.cc
utils/logalloc.cc label metrics with basic_level
2025-03-03 16:58:38 +02:00
logalloc.hh
utils: do not include unused headers
2025-01-14 07:56:39 -05:00
lru.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
managed_bytes.cc
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
managed_bytes.hh
managed_bytes: in the copy constructor, respect the target preferred allocation size
2025-04-16 22:06:06 +02:00
managed_ref.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
managed_vector.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
maybe_yield.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
memory_data_sink.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
memory_limit_reached.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
multiprecision_int.cc
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
multiprecision_int.hh
utils: do not include unused headers
2025-01-14 07:56:39 -05:00
murmur_hash.cc
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
murmur_hash.hh
utils: do not include unused headers
2025-01-14 07:56:39 -05:00
mutable_view.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
neat-object-id.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
observable.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
on_internal_error.cc
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
on_internal_error.hh
utils: do not include unused headers
2025-01-14 07:56:39 -05:00
only_on_shard0.hh
storage_service, tablets: Collect per-node capacity in load_stats
2025-03-06 12:17:32 +01:00
overloaded_functor.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
phased_barrier.hh
utils: phased_barrier, pluggable: use named gate
2025-04-12 11:47:00 +03:00
pluggable.hh
utils: phased_barrier, pluggable: use named gate
2025-04-12 11:47:00 +03:00
preempt.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
pretty_printers.cc
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
pretty_printers.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
rate_limiter.cc
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
rate_limiter.hh
utils: do not include unused headers
2025-01-14 07:56:39 -05:00
recent_entries_map.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
refcounted.hh
utils: add advanced_rpc_compressor
2024-12-23 23:37:02 +01:00
reservoir_sampling.hh
utils: introduce reservoir_sampling
2024-12-23 23:37:02 +01:00
result_combinators.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
result_loop.hh
utils: do not include unused headers
2025-01-14 07:56:39 -05:00
result_try.hh
utils: do not include unused headers
2025-01-14 07:56:39 -05:00
result.hh
utils: do not include unused headers
2025-01-14 07:56:39 -05:00
reusable_buffer.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
rjson.cc
utils: do not include unused headers
2025-01-14 07:56:39 -05:00
rjson.hh
utils/rjson: streaming_writer: add RawValue()
2025-02-17 08:01:38 -05:00
rpc_utils.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
runtime.cc
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
runtime.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
sequenced_set.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
sequential_producer.hh
Introduce LDAP role manager & saslauthd authenticator
2025-01-12 14:50:29 +02:00
serialization.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
serialized_action.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
shared_dict.hh
utils/hashers: add get_sha256()
2025-04-01 00:07:28 +02:00
simple_hashers.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
small_vector.hh
utils: small_vector: expose internal_capacity()
2024-12-24 12:19:20 +02:00
sorting.hh
utils/sorting: Fix VerticesContainer concept constraints
2025-02-26 23:23:53 +02:00
stall_free.hh
tree: correct misspellings
2025-01-26 15:54:06 +02:00
stream_compressor.cc
utils: correct misspellings
2025-01-02 16:47:57 +02:00
stream_compressor.hh
utils: correct misspellings
2025-01-02 16:47:57 +02:00
streaming_histogram.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
tagged_integer.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
throttle.hh
treewide: change assert() to SCYLLA_ASSERT()
2024-08-05 08:23:35 +03:00
to_string.cc
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
to_string.hh
utils: do not include unused headers
2025-01-14 07:56:39 -05:00
top_k.hh
utils: do not include unused headers
2025-01-14 07:56:39 -05:00
tuple_utils.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
unconst.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
unique_view.hh
utils: implement drop-in replacement for replacing boost::adaptors::uniqued
2025-01-21 16:24:45 +08:00
updateable_value.cc
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
updateable_value.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
user_provided_param.hh
utils: do not include unused headers
2025-01-14 07:56:39 -05:00
utf8.cc
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
utf8.hh
utils: do not include unused headers
2025-01-14 07:56:39 -05:00
UUID_gen.cc
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
UUID_gen.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
uuid.cc
tree: Remove unused boost headers
2025-02-25 10:32:32 +03:00
UUID.hh
utils: do not include unused headers
2025-01-14 07:56:39 -05:00
value_or_reference.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
variant_element.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
vle.hh
treewide: relicense to ScyllaDB-Source-Available-1.0
2024-12-18 17:45:13 +02:00
xx_hasher.hh
utils: do not include unused headers
2025-01-14 07:56:39 -05:00