memgraph/src/database/graph_db.hpp

145 lines
4.7 KiB
C++
Raw Normal View History

#pragma once
#include "cppitertools/filter.hpp"
#include "cppitertools/imap.hpp"
#include "data_structures/concurrent/concurrent_map.hpp"
#include "data_structures/concurrent/concurrent_set.hpp"
#include "data_structures/concurrent/skiplist.hpp"
#include "database/graph_db_datatypes.hpp"
#include "database/indexes/key_index.hpp"
#include "database/indexes/label_property_index.hpp"
Write-ahead log Summary: My dear fellow Memgraphians. It's friday afternoon, and I am as ready to pop as WAL is to get reviewed... What's done: - Vertices and Edges have global IDs, stored in `VersionList`. Main storage is now a concurrent map ID->vlist_ptr. - WriteAheadLog class added. It's based around buffering WAL::Op objects (elementraly DB changes) and periodically serializing and flusing them to disk. - Snapshot recovery refactored, WAL recovery added. Snapshot format changed again to include necessary info. - Durability testing completely reworked. What's not done (and should be when we decide how): - Old WAL file purging. - Config refactor (naming and organization). Will do when we discuss what we want. - Changelog and new feature documentation (both depending on the point above). - Better error handling and recovery feedback. Currently it's all returning bools, which is not fine-grained enough (neither for errors nor partial successes, also EOF is reported as a failure at the moment). - Moving the implementation of WAL stuff to .cpp where possible. - Not sure if there are transactions being created outside of `GraphDbAccessor` and it's `BuildIndex`. Need to look into. - True write-ahead logic (flag controlled): not committing a DB transaction if the WAL has not flushed it's data. We can discuss the gain/effort ratio for this feature. Reviewers: buda, mislav.bradac, teon.banek, dgleich Reviewed By: dgleich Subscribers: mtomic, pullbot Differential Revision: https://phabricator.memgraph.io/D958
2017-11-13 16:50:49 +08:00
#include "durability/wal.hpp"
#include "mvcc/version_list.hpp"
#include "storage/deferred_deleter.hpp"
#include "storage/edge.hpp"
#include "storage/garbage_collector.hpp"
#include "storage/unique_object_store.hpp"
#include "storage/vertex.hpp"
#include "transactions/engine.hpp"
#include "utils/scheduler.hpp"
/**
* Main class which represents Database concept in code.
* This class is essentially a data structure. It exposes
* all the data publicly, and should therefore not be directly
* exposed to client functions. The GraphDbAccessor is used for that.
*
* Always be sure that GraphDb object is destructed before main exits, i. e.
* GraphDb object shouldn't be part of global/static variable, except if its
* destructor is explicitly called before main exits. Consider code:
*
* GraphDb db; // KeyIndex is created as a part of db.
* int main() {
* GraphDbAccessor dba(db);
* auto v = dba.InsertVertex();
* v.add_label(dba.Label(
* "Start")); // New SkipList is created in KeyIndex for LabelIndex.
* // That SkipList creates SkipListGc which
* // initialises static Executor object.
* return 0;
* }
*
* After main exits: 1. Executor is destructed, 2. KeyIndex is destructed.
* Destructor of KeyIndex calls delete on created SkipLists which destroy
* SkipListGc that tries to use Excutioner object that doesn't exist anymore.
* -> CRASH
*/
class GraphDb {
public:
/// GraphDb configuration. Initialized from flags, but modifiable.
struct Config {
Config();
// Durability flags.
bool durability_enabled;
std::string durability_directory;
bool db_recover_on_startup;
int snapshot_cycle_sec;
int snapshot_max_retained;
int snapshot_on_exit;
// Misc flags.
int gc_cycle_sec;
int query_execution_time_sec;
};
explicit GraphDb(Config config = Config{});
/** Delete all vertices and edges and free all deferred deleters. */
~GraphDb();
/** Database object can't be copied. */
GraphDb(const GraphDb &db) = delete;
GraphDb(GraphDb &&other) = default;
GraphDb &operator=(const GraphDb &other) = default;
GraphDb &operator=(GraphDb &&other) = default;
/** Stop all transactions and set is_accepting_transactions_ to false. */
void Shutdown();
void CollectGarbage();
/** When this is false, no new transactions should be created. */
std::atomic<bool> is_accepting_transactions_{true};
private:
friend class GraphDbAccessor;
void StartSnapshooting();
Config config_;
/** transaction engine related to this database */
tx::Engine tx_engine_;
Write-ahead log Summary: My dear fellow Memgraphians. It's friday afternoon, and I am as ready to pop as WAL is to get reviewed... What's done: - Vertices and Edges have global IDs, stored in `VersionList`. Main storage is now a concurrent map ID->vlist_ptr. - WriteAheadLog class added. It's based around buffering WAL::Op objects (elementraly DB changes) and periodically serializing and flusing them to disk. - Snapshot recovery refactored, WAL recovery added. Snapshot format changed again to include necessary info. - Durability testing completely reworked. What's not done (and should be when we decide how): - Old WAL file purging. - Config refactor (naming and organization). Will do when we discuss what we want. - Changelog and new feature documentation (both depending on the point above). - Better error handling and recovery feedback. Currently it's all returning bools, which is not fine-grained enough (neither for errors nor partial successes, also EOF is reported as a failure at the moment). - Moving the implementation of WAL stuff to .cpp where possible. - Not sure if there are transactions being created outside of `GraphDbAccessor` and it's `BuildIndex`. Need to look into. - True write-ahead logic (flag controlled): not committing a DB transaction if the WAL has not flushed it's data. We can discuss the gain/effort ratio for this feature. Reviewers: buda, mislav.bradac, teon.banek, dgleich Reviewed By: dgleich Subscribers: mtomic, pullbot Differential Revision: https://phabricator.memgraph.io/D958
2017-11-13 16:50:49 +08:00
std::atomic<int64_t> next_vertex_id_{0};
std::atomic<int64_t> next_edge_id{0};
// main storage for the graph
Write-ahead log Summary: My dear fellow Memgraphians. It's friday afternoon, and I am as ready to pop as WAL is to get reviewed... What's done: - Vertices and Edges have global IDs, stored in `VersionList`. Main storage is now a concurrent map ID->vlist_ptr. - WriteAheadLog class added. It's based around buffering WAL::Op objects (elementraly DB changes) and periodically serializing and flusing them to disk. - Snapshot recovery refactored, WAL recovery added. Snapshot format changed again to include necessary info. - Durability testing completely reworked. What's not done (and should be when we decide how): - Old WAL file purging. - Config refactor (naming and organization). Will do when we discuss what we want. - Changelog and new feature documentation (both depending on the point above). - Better error handling and recovery feedback. Currently it's all returning bools, which is not fine-grained enough (neither for errors nor partial successes, also EOF is reported as a failure at the moment). - Moving the implementation of WAL stuff to .cpp where possible. - Not sure if there are transactions being created outside of `GraphDbAccessor` and it's `BuildIndex`. Need to look into. - True write-ahead logic (flag controlled): not committing a DB transaction if the WAL has not flushed it's data. We can discuss the gain/effort ratio for this feature. Reviewers: buda, mislav.bradac, teon.banek, dgleich Reviewed By: dgleich Subscribers: mtomic, pullbot Differential Revision: https://phabricator.memgraph.io/D958
2017-11-13 16:50:49 +08:00
ConcurrentMap<int64_t, mvcc::VersionList<Vertex> *> vertices_;
ConcurrentMap<int64_t, mvcc::VersionList<Edge> *> edges_;
// Garbage collectors
Write-ahead log Summary: My dear fellow Memgraphians. It's friday afternoon, and I am as ready to pop as WAL is to get reviewed... What's done: - Vertices and Edges have global IDs, stored in `VersionList`. Main storage is now a concurrent map ID->vlist_ptr. - WriteAheadLog class added. It's based around buffering WAL::Op objects (elementraly DB changes) and periodically serializing and flusing them to disk. - Snapshot recovery refactored, WAL recovery added. Snapshot format changed again to include necessary info. - Durability testing completely reworked. What's not done (and should be when we decide how): - Old WAL file purging. - Config refactor (naming and organization). Will do when we discuss what we want. - Changelog and new feature documentation (both depending on the point above). - Better error handling and recovery feedback. Currently it's all returning bools, which is not fine-grained enough (neither for errors nor partial successes, also EOF is reported as a failure at the moment). - Moving the implementation of WAL stuff to .cpp where possible. - Not sure if there are transactions being created outside of `GraphDbAccessor` and it's `BuildIndex`. Need to look into. - True write-ahead logic (flag controlled): not committing a DB transaction if the WAL has not flushed it's data. We can discuss the gain/effort ratio for this feature. Reviewers: buda, mislav.bradac, teon.banek, dgleich Reviewed By: dgleich Subscribers: mtomic, pullbot Differential Revision: https://phabricator.memgraph.io/D958
2017-11-13 16:50:49 +08:00
GarbageCollector<ConcurrentMap<int64_t, mvcc::VersionList<Vertex> *>, Vertex>
gc_vertices_;
GarbageCollector<ConcurrentMap<int64_t, mvcc::VersionList<Edge> *>, Edge>
gc_edges_;
// Deleters for not relevant records
DeferredDeleter<Vertex> vertex_record_deleter_;
DeferredDeleter<Edge> edge_record_deleter_;
// Deleters for not relevant version_lists
DeferredDeleter<mvcc::VersionList<Vertex>> vertex_version_list_deleter_;
DeferredDeleter<mvcc::VersionList<Edge>> edge_version_list_deleter_;
// unique object stores
// TODO this should be also garbage collected
ConcurrentSet<std::string> labels_;
ConcurrentSet<std::string> edge_types_;
ConcurrentSet<std::string> properties_;
// indexes
KeyIndex<GraphDbTypes::Label, Vertex> labels_index_;
LabelPropertyIndex label_property_index_;
/**
* Flag indicating if index building is in progress. Memgraph does not support
* concurrent index builds on the same database (transaction engine), so we
* reject index builds if there is one in progress. See
* GraphDbAccessor::BuildIndex.
*/
std::atomic<bool> index_build_in_progress_{false};
Write-ahead log Summary: My dear fellow Memgraphians. It's friday afternoon, and I am as ready to pop as WAL is to get reviewed... What's done: - Vertices and Edges have global IDs, stored in `VersionList`. Main storage is now a concurrent map ID->vlist_ptr. - WriteAheadLog class added. It's based around buffering WAL::Op objects (elementraly DB changes) and periodically serializing and flusing them to disk. - Snapshot recovery refactored, WAL recovery added. Snapshot format changed again to include necessary info. - Durability testing completely reworked. What's not done (and should be when we decide how): - Old WAL file purging. - Config refactor (naming and organization). Will do when we discuss what we want. - Changelog and new feature documentation (both depending on the point above). - Better error handling and recovery feedback. Currently it's all returning bools, which is not fine-grained enough (neither for errors nor partial successes, also EOF is reported as a failure at the moment). - Moving the implementation of WAL stuff to .cpp where possible. - Not sure if there are transactions being created outside of `GraphDbAccessor` and it's `BuildIndex`. Need to look into. - True write-ahead logic (flag controlled): not committing a DB transaction if the WAL has not flushed it's data. We can discuss the gain/effort ratio for this feature. Reviewers: buda, mislav.bradac, teon.banek, dgleich Reviewed By: dgleich Subscribers: mtomic, pullbot Differential Revision: https://phabricator.memgraph.io/D958
2017-11-13 16:50:49 +08:00
durability::WriteAheadLog wal_;
// Schedulers
Scheduler gc_scheduler_;
Scheduler snapshot_creator_;
// Periodically wakes up and hints to transactions that are running for a long
// time to stop their execution.
Scheduler transaction_killer_;
/** DB level global counters, used in the "counter" function. */
ConcurrentMap<std::string, std::atomic<int64_t>> counters_;
};