reconcilable_result_builder passes range tombstone changes to _rt_assembler
using table schema, not query schema.
This means that a tombstone with bounds (a; b), where a < b in query schema
but a > b in table schema, will not be emitted from mutation_query.
This is a very serious bug, because it means that such tombstones in reverse
queries are not reconciled with data from other replicas.
If *any* queried replica has a row, but not the range tombstone which deleted
the row, the reconciled result will contain the deleted row.
In particular, range deletes performed while a replica is down will not
later be visible to reverse queries which select this replica, regardless of the
consistency level.
As far as I can see, this doesn't result in any persistent data loss.
Only in that some data might appear resurrected to reverse queries,
until the relevant range tombstone is fully repaired.
This series fixes the bug and adds a minimal reproducer test.
Fixes #10598
Closes scylladb/scylladb#16003
* github.com:scylladb/scylladb:
mutation_query_test: test that range tombstones are sent in reverse queries
mutation_query: properly send range tombstones in reverse queries
(cherry picked from commit 65e42e4166)
184 lines
6.2 KiB
C++
184 lines
6.2 KiB
C++
/*
|
|
* Copyright (C) 2015-present ScyllaDB
|
|
*/
|
|
|
|
/*
|
|
* SPDX-License-Identifier: AGPL-3.0-or-later
|
|
*/
|
|
|
|
#pragma once
|
|
|
|
#include "query-request.hh"
|
|
#include "query-result.hh"
|
|
#include "mutation/frozen_mutation.hh"
|
|
#include "db/timeout_clock.hh"
|
|
#include "mutation/mutation.hh"
|
|
#include "utils/chunked_vector.hh"
|
|
#include "mutation/range_tombstone_assembler.hh"
|
|
|
|
class reconcilable_result;
|
|
class frozen_reconcilable_result;
|
|
class mutation_source;
|
|
|
|
// Can be read by other cores after publishing.
|
|
struct partition {
|
|
uint32_t _row_count_low_bits;
|
|
frozen_mutation _m; // FIXME: We don't need cf UUID, which frozen_mutation includes.
|
|
uint32_t _row_count_high_bits;
|
|
partition(uint32_t row_count_low_bits, frozen_mutation m, uint32_t row_count_high_bits)
|
|
: _row_count_low_bits(row_count_low_bits)
|
|
, _m(std::move(m))
|
|
, _row_count_high_bits(row_count_high_bits)
|
|
{ }
|
|
|
|
partition(uint64_t row_count, frozen_mutation m)
|
|
: _row_count_low_bits(static_cast<uint32_t>(row_count))
|
|
, _m(std::move(m))
|
|
, _row_count_high_bits(static_cast<uint32_t>(row_count >> 32))
|
|
{ }
|
|
|
|
uint32_t row_count_low_bits() const {
|
|
return _row_count_low_bits;
|
|
}
|
|
|
|
uint32_t row_count_high_bits() const {
|
|
return _row_count_high_bits;
|
|
}
|
|
|
|
uint64_t row_count() const {
|
|
return (static_cast<uint64_t>(_row_count_high_bits) << 32) | _row_count_low_bits;
|
|
}
|
|
|
|
const frozen_mutation& mut() const {
|
|
return _m;
|
|
}
|
|
|
|
frozen_mutation& mut() {
|
|
return _m;
|
|
}
|
|
|
|
|
|
bool operator==(const partition& other) const {
|
|
return row_count() == other.row_count() && _m.representation() == other._m.representation();
|
|
}
|
|
};
|
|
|
|
// The partitions held by this object are ordered according to dht::decorated_key ordering and non-overlapping.
|
|
// Each mutation must have different key.
|
|
//
|
|
// Can be read by other cores after publishing.
|
|
class reconcilable_result {
|
|
uint32_t _row_count_low_bits;
|
|
query::short_read _short_read;
|
|
query::result_memory_tracker _memory_tracker;
|
|
utils::chunked_vector<partition> _partitions;
|
|
uint32_t _row_count_high_bits;
|
|
public:
|
|
~reconcilable_result();
|
|
reconcilable_result();
|
|
reconcilable_result(reconcilable_result&&) = default;
|
|
reconcilable_result& operator=(reconcilable_result&&) = default;
|
|
reconcilable_result(uint32_t row_count_low_bits, utils::chunked_vector<partition> partitions, query::short_read short_read,
|
|
uint32_t row_count_high_bits, query::result_memory_tracker memory_tracker = { });
|
|
reconcilable_result(uint64_t row_count, utils::chunked_vector<partition> partitions, query::short_read short_read,
|
|
query::result_memory_tracker memory_tracker = { });
|
|
|
|
const utils::chunked_vector<partition>& partitions() const;
|
|
utils::chunked_vector<partition>& partitions();
|
|
|
|
uint32_t row_count_low_bits() const {
|
|
return _row_count_low_bits;
|
|
}
|
|
|
|
uint32_t row_count_high_bits() const {
|
|
return _row_count_high_bits;
|
|
}
|
|
|
|
uint64_t row_count() const {
|
|
return (static_cast<uint64_t>(_row_count_high_bits) << 32) | _row_count_low_bits;
|
|
}
|
|
|
|
query::short_read is_short_read() const {
|
|
return _short_read;
|
|
}
|
|
|
|
size_t memory_usage() const {
|
|
return _memory_tracker.used_memory();
|
|
}
|
|
|
|
bool operator==(const reconcilable_result& other) const;
|
|
|
|
struct printer {
|
|
const reconcilable_result& self;
|
|
schema_ptr schema;
|
|
friend std::ostream& operator<<(std::ostream&, const printer&);
|
|
};
|
|
|
|
printer pretty_printer(schema_ptr) const;
|
|
};
|
|
|
|
class reconcilable_result_builder {
|
|
const schema& _schema;
|
|
const query::partition_slice& _slice;
|
|
bool _reversed;
|
|
schema_ptr _query_schema;
|
|
|
|
bool _return_static_content_on_partition_with_no_rows{};
|
|
bool _static_row_is_alive{};
|
|
uint64_t _total_live_rows = 0;
|
|
query::result_memory_accounter _memory_accounter;
|
|
stop_iteration _stop;
|
|
std::optional<streamed_mutation_freezer> _mutation_consumer;
|
|
range_tombstone_assembler _rt_assembler;
|
|
|
|
uint64_t _live_rows{};
|
|
// make this the last member so it is destroyed first. #7240
|
|
utils::chunked_vector<partition> _result;
|
|
size_t _used_at_entry;
|
|
|
|
private:
|
|
stop_iteration consume(range_tombstone&& rt);
|
|
|
|
public:
|
|
// Expects table schema (non-reversed) and half-reversed (legacy) slice when building results for reverse query.
|
|
reconcilable_result_builder(const schema& s, const query::partition_slice& slice,
|
|
query::result_memory_accounter&& accounter) noexcept
|
|
: _schema(s), _slice(slice), _reversed(_slice.options.contains(query::partition_slice::option::reversed))
|
|
, _query_schema(_reversed ? _schema.make_reversed() : _schema.shared_from_this())
|
|
, _memory_accounter(std::move(accounter))
|
|
{ }
|
|
|
|
void consume_new_partition(const dht::decorated_key& dk);
|
|
void consume(tombstone t);
|
|
stop_iteration consume(static_row&& sr, tombstone, bool is_alive);
|
|
stop_iteration consume(clustering_row&& cr, row_tombstone, bool is_alive);
|
|
stop_iteration consume(range_tombstone_change&& rtc);
|
|
stop_iteration consume_end_of_partition();
|
|
reconcilable_result consume_end_of_stream();
|
|
};
|
|
|
|
future<query::result> to_data_query_result(
|
|
const reconcilable_result&,
|
|
schema_ptr,
|
|
const query::partition_slice&,
|
|
uint64_t row_limit,
|
|
uint32_t partition_limit,
|
|
query::result_options opts = query::result_options::only_result());
|
|
|
|
// Query the content of the mutation.
|
|
//
|
|
// The mutation is destroyed in the process, see `mutation::consume()`.
|
|
query::result query_mutation(
|
|
mutation&& m,
|
|
const query::partition_slice& slice,
|
|
uint64_t row_limit = query::max_rows,
|
|
gc_clock::time_point now = gc_clock::now(),
|
|
query::result_options opts = query::result_options::only_result());
|
|
|
|
// Performs a query for counter updates.
|
|
future<mutation_opt> counter_write_query(schema_ptr, const mutation_source&, reader_permit permit,
|
|
const dht::decorated_key& dk,
|
|
const query::partition_slice& slice,
|
|
tracing::trace_state_ptr trace_ptr);
|
|
|