database/mariadb-columnstore-engine
1
0
Fork 0
mirror of https://github.com/mariadb-corporation/mariadb-columnstore-engine.git synced 2025-04-18 21:44:02 +03:00
Code
mariadb-columnstore-engine/versioning/BRM/extentmap.h
Aleksei Antipovskii 5556d818f8 chore(codestyle): mark virtual methods as override
2025-02-21 20:02:38 +04:00

1231 lines
47 KiB
C++
Raw Permalink Blame History

/* Copyright (C) 2014 InfiniDB, Inc.
Copyright (C) 2016-2022 MariaDB Corporation
This program is free software; you can redistribute it and/or
modify it under the terms of the GNU General Public License
as published by the Free Software Foundation; version 2 of
the License.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program; if not, write to the Free Software
Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston,
MA 02110-1301, USA. */
/******************************************************************************
* $Id: extentmap.h 1936 2013-07-09 22:10:29Z dhall $
*
*****************************************************************************/
/** @file
* class ExtentMap
*/
#pragma once
#include <sys/types.h>
#include <vector>
#include <set>
#include <unordered_map>
#include <tr1/unordered_map>
#include <mutex>
// #define NDEBUG
#include <cassert>
#include <boost/functional/hash.hpp> //boost::hash
#include <boost/interprocess/allocators/allocator.hpp>
#include <boost/interprocess/containers/vector.hpp>
#include <boost/interprocess/containers/map.hpp>
#include <boost/interprocess/managed_shared_memory.hpp>
#include <boost/interprocess/mapped_region.hpp>
#include <boost/interprocess/shared_memory_object.hpp>
#include <boost/unordered_map.hpp>
#include "shmkeys.h"
#include "brmtypes.h"
#include "mastersegmenttable.h"
#include "undoable.h"
#include "brmshmimpl.h"
#include "exceptclasses.h"
#ifdef NONE
#undef NONE
#endif
#ifdef READ
#undef READ
#endif
#ifdef WRITE
#undef WRITE
#endif
#define EXPORT
namespace bi = boost::interprocess;
namespace oam
{
typedef std::vector<uint16_t> DBRootConfigList;
}
namespace idbdatafile
{
class IDBDataFile;
}
namespace BRM
{
#define EM_MAGIC_V1 0x76f78b1c
#define EM_MAGIC_V2 0x76f78b1d
#define EM_MAGIC_V3 0x76f78b1e
#define EM_MAGIC_V4 0x76f78b1f
#define EM_MAGIC_V5 0x76f78b20
using PartitionNumberT = uint32_t;
using DBRootT = uint16_t;
using SegmentT = uint16_t;
using LastExtentIndexT = int;
using EmptyEMEntry = int;
using HighestOffset = uint32_t;
using LastIndEmptyIndEmptyInd = std::pair<LastExtentIndexT, EmptyEMEntry>;
using DBRootVec = std::vector<DBRootT>;
// assumed column width when calculating dictionary store extent size
#define DICT_COL_WIDTH 8
// valid values for EMEntry.status
const int16_t EXTENTSTATUSMIN(0); // equal to minimum valid status value
const int16_t EXTENTAVAILABLE(0);
const int16_t EXTENTUNAVAILABLE(1);
const int16_t EXTENTOUTOFSERVICE(2);
const int16_t EXTENTSTATUSMAX(2); // equal to maximum valid status value
enum partition_type_enum
{
PART_DEFAULT = 0,
PART_CASUAL,
PART_RANGE
};
typedef partition_type_enum EMPartitionType_t;
typedef int64_t RangePartitionData_t;
const char CP_INVALID = 0;
const char CP_UPDATING = 1;
const char CP_VALID = 2;
struct EMCasualPartition_struct_v4
{
RangePartitionData_t hi_val; // This needs to be reinterpreted as unsigned for uint64_t column types.
RangePartitionData_t lo_val;
int32_t sequenceNum;
char isValid; // CP_INVALID - No min/max and no DML in progress. CP_UPDATING - Update in progress.
// CP_VALID- min/max is valid
};
struct EMPartition_struct_v4
{
EMCasualPartition_struct_v4 cprange;
};
struct EMEntry_v4
{
InlineLBIDRange range;
int fileID;
uint32_t blockOffset;
HWM_t HWM;
PartitionNumberT partitionNum; // starts at 0
uint16_t segmentNum; // starts at 0
DBRootT dbRoot; // starts at 1 to match Columnstore.xml
uint16_t colWid;
int16_t status; // extent avail for query or not, or out of service
EMPartition_struct_v4 partition;
};
// MCOL-641: v5 structs of the extent map. This version supports int128_t min
// and max values for casual partitioning.
struct EMCasualPartition_struct
{
int32_t sequenceNum;
char isValid; // CP_INVALID - No min/max and no DML in progress. CP_UPDATING - Update in progress.
// CP_VALID- min/max is valid
union
{
int128_t bigLoVal; // These need to be reinterpreted as unsigned for uint64_t/uint128_t column types.
int64_t loVal;
};
union
{
int128_t bigHiVal;
int64_t hiVal;
};
EXPORT EMCasualPartition_struct();
EXPORT EMCasualPartition_struct(int64_t lo, int64_t hi, int32_t seqNum);
EXPORT EMCasualPartition_struct(const int128_t bigLo, const int128_t bigHi, int32_t seqNum);
EXPORT EMCasualPartition_struct(int64_t lo, int64_t hi, int32_t seqNum, char status);
EXPORT EMCasualPartition_struct(const EMCasualPartition_struct& em);
EXPORT EMCasualPartition_struct& operator=(const EMCasualPartition_struct& em);
};
using EMCasualPartition_t = EMCasualPartition_struct;
struct EMPartition_struct
{
EMCasualPartition_t cprange;
};
typedef EMPartition_struct EMPartition_t;
struct EMEntry
{
InlineLBIDRange range;
int fileID;
uint32_t blockOffset;
HWM_t HWM;
PartitionNumberT partitionNum; // starts at 0
uint16_t segmentNum; // starts at 0
DBRootT dbRoot; // starts at 1 to match Columnstore.xml
uint16_t colWid;
int16_t status; // extent avail for query or not, or out of service
EMPartition_t partition;
EXPORT EMEntry();
EMEntry(const InlineLBIDRange& range, int fileID, uint32_t bOffset, HWM_t hwm, PartitionNumberT pNum,
uint16_t sNum, DBRootT dRoot, uint16_t cWid, int16_t s, EMPartition_t partition);
EXPORT EMEntry(const EMEntry&);
EXPORT EMEntry& operator=(const EMEntry&);
EXPORT bool operator<(const EMEntry&) const;
};
// Bug 2989, moved from joblist
struct ExtentSorter
{
bool operator()(const EMEntry& e1, const EMEntry& e2)
{
if (e1.dbRoot < e2.dbRoot)
return true;
if (e1.dbRoot == e2.dbRoot && e1.partitionNum < e2.partitionNum)
return true;
if (e1.dbRoot == e2.dbRoot && e1.partitionNum == e2.partitionNum && e1.blockOffset < e2.blockOffset)
return true;
if (e1.dbRoot == e2.dbRoot && e1.partitionNum == e2.partitionNum && e1.blockOffset == e2.blockOffset &&
e1.segmentNum < e2.segmentNum)
return true;
return false;
}
};
using ShmSegmentManagerT = bi::managed_shared_memory::segment_manager;
using ShmVoidAllocator = bi::allocator<void, ShmSegmentManagerT>;
using LBID_tAlloc = bi::allocator<LBID_t, ShmSegmentManagerT>;
using PartitionNumberTAlloc = bi::allocator<PartitionNumberT, ShmSegmentManagerT>;
using LBID_tVectorT = bi::vector<LBID_t, LBID_tAlloc>;
using PartitionIndexContainerKeyT = PartitionNumberT;
using PartitionIndexContainerValT = std::pair<const PartitionIndexContainerKeyT, LBID_tVectorT>;
using PartitionIndexContainerValTAlloc = bi::allocator<PartitionIndexContainerValT, ShmSegmentManagerT>;
// Can't use std::unordered_map presumably b/c the map's pointer type doesn't use offset_type as boost::u_map
// does
using PartitionIndexContainerT =
boost::unordered_map<PartitionIndexContainerKeyT, LBID_tVectorT, boost::hash<PartitionIndexContainerKeyT>,
std::equal_to<PartitionIndexContainerKeyT>, PartitionIndexContainerValTAlloc>;
using OIDIndexContainerKeyT = OID_t;
using OIDIndexContainerValT = std::pair<const OIDIndexContainerKeyT, PartitionIndexContainerT>;
using OIDIndexContainerValTAlloc = bi::allocator<OIDIndexContainerValT, ShmSegmentManagerT>;
using OIDIndexContainerT =
boost::unordered_map<OIDIndexContainerKeyT, PartitionIndexContainerT, boost::hash<OIDIndexContainerKeyT>,
std::equal_to<OIDIndexContainerKeyT>, OIDIndexContainerValTAlloc>;
using DBRootIndexTAlloc = bi::allocator<OIDIndexContainerT, ShmSegmentManagerT>;
using DBRootIndexContainerT = bi::vector<OIDIndexContainerT, DBRootIndexTAlloc>;
using ExtentMapIndex = DBRootIndexContainerT;
using LBID_tFindResult = bi::vector<LBID_t>;
using InsertUpdateShmemKeyPair = std::pair<bool, bool>;
// RBTREE.
using EMEntryKeyValueType = std::pair<const int64_t, EMEntry>;
using VoidAllocator =
boost::interprocess::allocator<void, boost::interprocess::managed_shared_memory::segment_manager>;
using EMEntryKeyValueTypeAllocator =
boost::interprocess::allocator<EMEntryKeyValueType,
boost::interprocess::managed_shared_memory::segment_manager>;
using ExtentMapRBTree =
boost::interprocess::map<int64_t, EMEntry, std::less<int64_t>, EMEntryKeyValueTypeAllocator>;
class ExtentMapRBTreeImpl
{
public:
~ExtentMapRBTreeImpl() = default;
static ExtentMapRBTreeImpl* makeExtentMapRBTreeImpl(unsigned key, off_t size, bool readOnly = false);
static void refreshShmWithLock()
{
boost::mutex::scoped_lock lk(fInstanceMutex);
return refreshShm();
}
static void refreshShm()
{
if (fInstance)
{
delete fInstance;
fInstance = nullptr;
}
}
inline void grow(unsigned key, off_t size)
{
fManagedShm.grow(key, size);
}
inline unsigned key() const
{
return fManagedShm.key();
}
inline ExtentMapRBTree* get() const
{
VoidAllocator allocator(fManagedShm.fShmSegment->get_segment_manager());
return fManagedShm.fShmSegment->find_or_construct<ExtentMapRBTree>("EmMapRBTree")(std::less<LBID_t>(),
allocator);
}
inline uint64_t getFreeMemory() const
{
return fManagedShm.fShmSegment->get_free_memory();
}
inline uint64_t getSize() const
{
return fManagedShm.fShmSegment->get_size();
}
private:
ExtentMapRBTreeImpl(unsigned key, off_t size, bool readOnly = false);
ExtentMapRBTreeImpl(const ExtentMapRBTreeImpl& rhs);
ExtentMapRBTreeImpl& operator=(const ExtentMapRBTreeImpl& rhs);
BRMManagedShmImplRBTree fManagedShm;
static boost::mutex fInstanceMutex;
static ExtentMapRBTreeImpl* fInstance;
};
class FreeListImpl
{
public:
~FreeListImpl() = default;
static FreeListImpl* makeFreeListImpl(unsigned key, off_t size, bool readOnly = false);
static void refreshShmWithLock()
{
boost::mutex::scoped_lock lk(fInstanceMutex);
return refreshShm();
}
static void refreshShm()
{
if (fInstance)
{
delete fInstance;
fInstance = nullptr;
}
}
inline void grow(unsigned key, off_t size)
#ifdef NDEBUG
{
fFreeList.grow(key, size);
}
#else
{
int rc = fFreeList.grow(key, size);
idbassert(rc == 0);
}
#endif
inline void makeReadOnly()
{
fFreeList.setReadOnly();
}
inline void clear(unsigned key, off_t size)
{
fFreeList.clear(key, size);
}
inline void swapout(BRMShmImpl& rhs)
{
fFreeList.swap(rhs);
rhs.destroy();
}
inline unsigned key() const
{
return fFreeList.key();
}
inline InlineLBIDRange* get() const
{
return reinterpret_cast<InlineLBIDRange*>(fFreeList.fMapreg.get_address());
}
private:
FreeListImpl(unsigned key, off_t size, bool readOnly = false);
FreeListImpl(const FreeListImpl& rhs);
FreeListImpl& operator=(const FreeListImpl& rhs);
BRMShmImpl fFreeList;
static boost::mutex fInstanceMutex;
static FreeListImpl* fInstance;
};
class ExtentMapIndexImpl
{
public:
~ExtentMapIndexImpl() = default;
static ExtentMapIndexImpl* makeExtentMapIndexImpl(unsigned key, off_t size, bool readOnly = false);
static void refreshShm()
{
if (fInstance_)
{
delete fInstance_;
fInstance_ = nullptr;
}
}
// The multipliers and constants here are pure theoretical
// tested using customer's data.
static size_t estimateEMIndexSize(uint32_t numberOfExtents)
{
// These are just educated guess values to calculate initial
// managed shmem size.
constexpr const size_t tablesNumber_ = 100ULL;
constexpr const size_t columnsNumber_ = 200ULL;
constexpr const size_t dbRootsNumber_ = 3ULL;
constexpr const size_t filesInPartition_ = 4ULL;
constexpr const size_t extentsInPartition_ = filesInPartition_ * 2;
return numberOfExtents * emIdentUnitSize_ +
numberOfExtents / extentsInPartition_ * partitionContainerUnitSize_ +
dbRootsNumber_ * tablesNumber_ * columnsNumber_;
}
bool growIfNeeded(const size_t memoryNeeded);
inline void grow(off_t size)
{
int rc = fBRMManagedShmMemImpl_.grow(size);
idbassert(rc == 0);
}
// After this call one needs to refresh any refs or ptrs sourced
// from this shmem.
inline void makeReadOnly()
{
fBRMManagedShmMemImpl_.setReadOnly();
}
inline void swapout(BRMManagedShmImpl& rhs)
{
fBRMManagedShmMemImpl_.swap(rhs);
}
inline unsigned key() const
{
return fBRMManagedShmMemImpl_.key();
}
unsigned getShmemSize()
{
return fBRMManagedShmMemImpl_.getManagedSegment()->get_size();
}
size_t getShmemFree()
{
return fBRMManagedShmMemImpl_.getManagedSegment()->get_free_memory();
}
unsigned getShmemImplSize()
{
return fBRMManagedShmMemImpl_.size();
}
void createExtentMapIndexIfNeeded();
ExtentMapIndex* get();
InsertUpdateShmemKeyPair insert(const EMEntry& emEntry, const LBID_t lbid);
InsertUpdateShmemKeyPair insert2ndLayerWrapper(OIDIndexContainerT& oids, const EMEntry& emEntry,
const LBID_t lbid, const bool aShmemHasGrown);
InsertUpdateShmemKeyPair insert2ndLayer(OIDIndexContainerT& oids, const EMEntry& emEntry, const LBID_t lbid,
const bool aShmemHasGrown);
InsertUpdateShmemKeyPair insert3dLayerWrapper(PartitionIndexContainerT& partitions, const EMEntry& emEntry,
const LBID_t lbid, const bool aShmemHasGrown);
InsertUpdateShmemKeyPair insert3dLayer(PartitionIndexContainerT& partitions, const EMEntry& emEntry,
const LBID_t lbid, const bool aShmemHasGrown);
LBID_tFindResult find(const DBRootT dbroot, const OID_t oid, const PartitionNumberT partitionNumber);
LBID_tFindResult find(const DBRootT dbroot, const OID_t oid);
LBID_tFindResult search2ndLayer(OIDIndexContainerT& oids, const OID_t oid,
const PartitionNumberT partitionNumber);
LBID_tFindResult search2ndLayer(OIDIndexContainerT& oids, const OID_t oid);
LBID_tFindResult search3dLayer(PartitionIndexContainerT& partitions,
const PartitionNumberT partitionNumber);
bool isDBRootEmpty(const DBRootT dbroot);
void deleteDbRoot(const DBRootT dbroot);
void deleteOID(const DBRootT dbroot, const OID_t oid);
void deleteEMEntry(const EMEntry& emEntry, const LBID_t lbid);
private:
BRMManagedShmImpl fBRMManagedShmMemImpl_;
ExtentMapIndexImpl(unsigned key, off_t size, bool readOnly = false);
ExtentMapIndexImpl(const ExtentMapIndexImpl& rhs);
ExtentMapIndexImpl& operator=(const ExtentMapIndexImpl& rhs);
static std::mutex fInstanceMutex_;
static ExtentMapIndexImpl* fInstance_;
static const constexpr uint32_t dbRootContainerUnitSize_ = 64ULL;
static const constexpr uint32_t oidContainerUnitSize_ = 352ULL; // 2 * map overhead
static const constexpr uint32_t partitionContainerUnitSize_ = 368ULL; // single map overhead
static const constexpr uint32_t emIdentUnitSize_ = sizeof(uint64_t);
static const constexpr uint32_t extraUnits_ = 2;
static const constexpr size_t freeSpaceThreshold_ = 256 * 1024;
};
/** @brief This class encapsulates the extent map functionality of the system
*
* This class encapsulates the extent map functionality of the system. It
* is currently implemented in the quickest-to-write (aka dumb) way to
* get something working into the hands of the other developers ASAP.
* The Extent Map shared data should be implemented in a more scalable
* structure such as a tree or hash table.
*/
class ExtentMap : public Undoable
{
public:
EXPORT ExtentMap();
EXPORT ~ExtentMap() override;
/** @brief Loads the ExtentMap entries from a file
*
* Loads the ExtentMap entries from a file. This will
* clear out any existing entries. The intention is that before
* the system starts, an external tool instantiates a single Extent
* Map and loads the stored entries.
* @param filename The file to load from.
* @note Throws an ios_base::failure exception on an IO error, runtime_error
* if the file "looks" bad.
*/
EXPORT void load(const std::string& filename, bool fixFL = false);
/** @brief Loads the ExtentMap entries from a binary blob.
*
* Loads the ExtentMap entries from a file. This will
* clear out any existing entries. The intention is that before
* the system starts, an external tool instantiates a single Extent
* Map and loads the stored entries.
* @param pointer to a binary blob.
*/
EXPORT void loadFromBinaryBlob(const char* blob);
/** @brief Saves the ExtentMap entries to a file
*
* Saves the ExtentMap entries to a file.
* @param filename The file to save to.
*/
EXPORT void save(const std::string& filename);
// @bug 1509. Added new version of lookup below.
/** @brief Returns the first and last LBID in the range for a given LBID
*
* Get the first and last LBID for the extent that contains the given LBID.
* @param LBID (in) The lbid to search for
* @param firstLBID (out) The first lbid for the extent
* @param lastLBID (out) the last lbid for the extent
* @return 0 on success, -1 on error
*/
EXPORT int lookup(LBID_t LBID, LBID_t& firstLBID, LBID_t& lastLBID);
// @bug 1055+. New functions added for multiple files per OID enhancement.
/** @brief Look up the OID and file block offset assiciated with an LBID
*
* Look up the OID and file block offset assiciated with an LBID
* @param LBID (in) The lbid to search for
* @param OID (out) The OID associated with lbid
* @param dbRoot (out) The db root containing the LBID
* @param partitionNum (out) The partition containing the LBID
* @param segmentNum (out) The segment containing the LBID
* @param fileBlockOffset (out) The file block offset associated
* with LBID
* @return 0 on success, -1 on error
*/
EXPORT int lookupLocal(LBID_t LBID, int& OID, uint16_t& dbRoot, uint32_t& partitionNum,
uint16_t& segmentNum, uint32_t& fileBlockOffset);
/** @brief Look up the LBID associated with a given OID, offset, partition, and segment.
*
* Look up the LBID associated with a given OID, offset, partition, and segment.
* @param OID (in) The OID to look up
* @param fileBlockOffset (in) The file block offset
* @param partitionNum (in) The partition containing the lbid
* @param segmentNum (in) The segement containing the lbid
* @param LBID (out) The LBID associated with the given offset of the OID.
* @return 0 on success, -1 on error
*/
EXPORT int lookupLocal(int OID, uint32_t partitionNum, uint16_t segmentNum, uint32_t fileBlockOffset,
LBID_t& LBID);
/** @brief Look up the LBID associated with a given dbroot, OID, offset,
* partition, and segment.
*
* Look up LBID associated with a given OID, offset, partition, and segment.
* @param OID (in) The OID to look up
* @param fileBlockOffset (in) The file block offset
* @param partitionNum (in) The partition containing the lbid
* @param segmentNum (in) The segement containing the lbid
* @param LBID (out) The LBID associated with the given offset of the OID.
* @return 0 on success, -1 on error
*/
EXPORT int lookupLocal_DBroot(int OID, uint16_t dbroot, uint32_t partitionNum, uint16_t segmentNum,
uint32_t fileBlockOffset, LBID_t& LBID);
// @bug 1055-.
/** @brief Look up the starting LBID associated with a given OID,
* partition, segment, and offset.
*
* @param OID (in) The OID to look up
* @param partitionNum (in) The partition containing the lbid
* @param segmentNum (in) The segement containing the lbid
* @param fileBlockOffset (in) The file block offset
* @param LBID (out) The starting LBID associated with the extent
* containing the given offset
* @return 0 on success, -1 on error
*/
int lookupLocalStartLbid(int OID, uint32_t partitionNum, uint16_t segmentNum, uint32_t fileBlockOffset,
LBID_t& LBID);
/** @brief Get a complete list of LBID ranges assigned to an OID
*
* Get a complete list of LBID ranges assigned to an OID.
*/
EXPORT void lookup(OID_t oid, LBIDRange_v& ranges);
/** @brief Allocate a "stripe" of extents for columns in a table (in DBRoot)
*
* If this is the first extent for the OID/DBRoot, it will start at
* file offset 0. If space for the OID already exists, the new
* extent will "logically" be appended to the end of the already-
* allocated space, although the extent may reside in a different
* physical file as indicated by dbRoot, partition, and segment.
* Partition and segment numbers are 0 based, dbRoot is 1 based.
*
* Allocate a "stripe" of extents for the specified columns and DBRoot
* @param cols (in) List of column OIDs and column widths
* @param dbRoot (in) DBRoot for requested extents.
* @param partitionNum (in/out) Partition number in file path.
* If allocating OID's first extent for this DBRoot, then
* partitionNum is input, else it is an output arg.
* @param segmentNum (out) Segment number selected for new extents.
* @param extents (out) list of lbids, numBlks, and fbo for new extents
* @return 0 on success, -1 on error
*/
EXPORT void createStripeColumnExtents(const std::vector<CreateStripeColumnExtentsArgIn>& cols,
uint16_t dbRoot, uint32_t& partitionNum, uint16_t& segmentNum,
std::vector<CreateStripeColumnExtentsArgOut>& extents);
/** @brief Allocates an extent for a column file
*
* Allocates an extent for the specified OID and DBroot.
* If this is the first extent for the OID/DBRoot, it will start at
* file offset 0. If space for the OID already exists, the new
* extent will "logically" be appended to the end of the already-
* allocated space, although the extent may reside in a different
* physical file as indicated by dbRoot, partition, and segment.
* Partition and segment numbers are 0 based, dbRoot is 1 based.
*
* @param OID (in) The OID requesting the extent.
* @param colWidth (in) Column width of the OID.
* @param dbRoot (in) DBRoot where extent is to be added.
* @param colDataType (in) the column type
* @param partitionNum (in/out) Partition number in file path.
* If allocating OID's first extent for this DBRoot, then
* partitionNum is input, else it is an output arg.
* @param segmentNum (out) Segment number assigned to the extent.
* @param lbid (out) The first LBID of the extent created.
* @param allocdsize (out) The total number of LBIDs allocated.
* @param startBlockOffset (out) The first block of the extent created.
* @param useLock Grab ExtentMap and FreeList WRITE lock to perform work
*/
// @bug 4091: To be deprecated as public function. Should just be a
// private function used by createStripeColumnExtents().
EXPORT void createColumnExtent_DBroot(int OID, uint32_t colWidth, uint16_t dbRoot,
execplan::CalpontSystemCatalog::ColDataType colDataType,
uint32_t& partitionNum, uint16_t& segmentNum, LBID_t& lbid,
int& allocdsize, uint32_t& startBlockOffset, bool useLock = true);
/** @brief Allocates extent for exact file that is specified
*
* Allocates an extent for the exact file specified by OID, DBRoot,
* partition, and segment.
* If this is the first extent for the OID/DBRoot, it will start at
* file offset 0. If space for the OID already exists, the new
* extent will "logically" be appended to the end of the already-
* allocated space.
* Partition and segment numbers are 0 based, dbRoot is 1 based.
*
* @param OID (in) The OID requesting the extent.
* @param colWidth (in) Column width of the OID.
* @param dbRoot (in) DBRoot where extent is to be added.
* @param partitionNum (in) Partition number in file path.
* If allocating OID's first extent for this DBRoot, then
* partitionNum is input, else it is an output arg.
* @param segmentNum (in) Segment number in file path.
* If allocating OID's first extent for this DBRoot, then
* segmentNum is input, else it is an output arg.
* @param colDataType (in) the column type
* @param lbid (out) The first LBID of the extent created.
* @param allocdSize (out) The total number of LBIDs allocated.
* @param startBlockOffset (out) The first block of the extent created.
*/
EXPORT void createColumnExtentExactFile(int OID, uint32_t colWidth, uint16_t dbRoot, uint32_t partitionNum,
uint16_t segmentNum,
execplan::CalpontSystemCatalog::ColDataType colDataType,
LBID_t& lbid, int& allocdsize, uint32_t& startBlockOffset);
/** @brief Allocates an extent for a dictionary store file
*
* Allocates an extent for the specified dictionary store OID,
* dbRoot, partition number, and segment number. These should
* correlate with those belonging to the corresponding token file.
* The first extent for each store file will start at file offset 0.
* Other extents will be appended to the end of the already-
* allocated space for the same store file.
* Partition and segment numbers are 0 based, dbRoot is 1 based.
*
* @param OID (in) The OID requesting the extent.
* @param dbRoot (in) DBRoot to assign to the extent.
* @param partitionNum (in) Partition number to assign to the extent.
* @param segmentNum (in) Segment number to assign to the extent.
* @param lbid (out) The first LBID of the extent created.
* @param allocdsize (out) The total number of LBIDs allocated.
*/
EXPORT void createDictStoreExtent(int OID, uint16_t dbRoot, uint32_t partitionNum, uint16_t segmentNum,
LBID_t& lbid, int& allocdsize);
/** @brief Rollback (delete) a set of extents for the specified OID.
*
* Deletes all the extents that logically follow the specified
* column extent; and sets the HWM for the specified extent.
* @param oid OID of the extents to be deleted.
* @param partitionNum Last partition to be kept.
* @param segmentNum Last segment in partitionNum to be kept.
* @param hwm HWM to be assigned to the last extent that is kept.
*/
EXPORT void rollbackColumnExtents(int oid, uint32_t partitionNum, uint16_t segmentNum, HWM_t hwm);
/** @brief Rollback (delete) set of extents for specified OID & DBRoot.
*
* Deletes all the extents that logically follow the specified
* column extent; and sets the HWM for the specified extent.
* @param oid OID of the extents to be deleted.
* @param bDeleteAll Flag indicates if all extents for oid and dbroot are
* to be deleted, else part#, seg#, and HWM are used.
* @param dbRoot DBRoot of the extents to be deleted.
* @param partitionNum Last partition to be kept.
* @param segmentNum Last segment in partitionNum to be kept.
* @param hwm HWM to be assigned to the last extent that is kept.
*/
EXPORT void rollbackColumnExtents_DBroot(int oid, bool bDeleteAll, uint16_t dbRoot, uint32_t partitionNum,
uint16_t segmentNum, HWM_t hwm);
/** @brief delete of column extents for the specified extents.
*
* Deletes the extents that logically follow the specified
* column extent in extentsInfo. It use the same algorithm as in
* rollbackColumnExtents.
* @param extentInfo the information for extents
*/
EXPORT void deleteEmptyColExtents(const ExtentsInfoMap_t& extentsInfo);
/** @brief delete of dictionary extents for the specified extents.
*
* Arguments specify the last stripe for all the oids. Any extents after this are
* deleted. The hwm's of the extents in the last stripe are updated
* based on the hwm in extentsInfo. It use the same algorithm as in
* rollbackDictStoreExtents.
* @param extentInfo the information for extents to be resetted
*/
EXPORT void deleteEmptyDictStoreExtents(const ExtentsInfoMap_t& extentsInfo);
/** @brief Rollback (delete) a set of dict store extents for an OID.
*
* Arguments specify the last stripe. Any extents after this are
* deleted. The hwm's of the extents in the last stripe are updated
* based on the contents of the hwm vector. If hwms is a partial list,
* (as in the first stripe of a partition), then any extents in sub-
* sequent segment files for that partition are deleted.
* @param oid OID of the extents to be deleted or updated.
* @param partitionNum Last partition to be kept.
* @param hwms Vector of hwms for the last partition to be kept.
*/
EXPORT void rollbackDictStoreExtents(int oid, uint32_t partitionNum, const std::vector<HWM_t>& hwms);
/** @brief Rollback (delete) a set of dict store extents for an OID & DBRoot
*
* Arguments specify the last stripe. Any extents after this are
* deleted. The hwm's of the extents in the last stripe are updated
* based on the contents of the hwm vector. If hwms is a partial list,
* (as in the first stripe of a partition), then any extents in sub-
* sequent segment files for that partition are deleted. If hwms is empty
* then all the extents in dbRoot are deleted.
* @param oid OID of the extents to be deleted or updated.
* @param dbRoot DBRoot of the extents to be deleted.
* @param partitionNum Last partition to be kept.
* @param segNums Vector of segment files in last partition to be kept.
* @param hwms Vector of hwms for the last partition to be kept.
*/
EXPORT void rollbackDictStoreExtents_DBroot(int oid, uint16_t dbRoot, uint32_t partitionNum,
const std::vector<uint16_t>& segNums,
const std::vector<HWM_t>& hwms);
/** @brief Deallocates all extents associated with OID
*
* Deallocates all extents associated with OID
* @param OID The OID to delete
*/
EXPORT void deleteOID(int OID);
/** @brief Deallocates all extents associated with each OID
*
* Deallocates all extents associated with each OID
* @param OIDs The OIDs to delete
*/
EXPORT void deleteOIDs(const OidsMap_t& OIDs);
/** @brief Check if any of the given partitions is the last one of a DBroot
*
* This is for partitioning operations to use. The last partition of a DBroot
* can not be dropped or disabled.
*
* @param OID (in) The OID
* @param partitionNums (in) The logical partition numbers to check.
* @return true if any of the partitions in the set is the last partition of
* a DBroot.
*/
/** @brief Gets the last local high water mark of an OID for a given dbRoot
*
* Get last local high water mark of an OID for a given dbRoot, relative to
* a segment file. The partition and segment numbers for the pertinent
* segment are also returned.
* @param OID (in) The OID
* @param dbRoot (in) The relevant DBRoot
* @param partitionNum (out) The relevant partition number
* @param segmentNum (out) The relevant segment number
* @param status (out) State of the extent (Available, OutOfService, etc)
* @param bFound (out) Indicates whether an extent was found for dbRoot
* @return The last file block number written to in the last
* partition/segment file for the given OID.
*/
EXPORT HWM_t getLastHWM_DBroot(int OID, uint16_t dbRoot, uint32_t& partitionNum, uint16_t& segmentNum,
int& status, bool& bFound);
/** @brief Gets the current high water mark of an OID,partition,segment
*
* Get current local high water mark of an OID, partition, segment;
* where HWM is relative to the specific segment file.
* @param OID (in) The OID
* @param partitionNum (in) The relevant partition number
* @param segmentNum (in) The relevant segment number
* @param status (out) State of the extent (Available, OutOfService, etc)
* @return The last file block number written to in the specified
* partition/segment file for the given OID.
*/
EXPORT HWM_t getLocalHWM(int OID, uint32_t partitionNum, uint16_t segmentNum, int& status);
/** @brief Sets the current high water mark of an OID,partition,segment
*
* Sets the current local high water mark of an OID, partition, segment;
* where HWM is relative to the specific segment file.
* @param OID The OID
* @param partitionNum (in) The relevant partition number
* @param segmentNum (in) The relevant segment number
* @param HWM The high water mark to record
*/
EXPORT void setLocalHWM(int OID, uint32_t partitionNum, uint16_t segmentNum, HWM_t HWM, bool firstNode,
bool uselock = true);
EXPORT void bulkSetHWM(const std::vector<BulkSetHWMArg>&, bool firstNode);
EXPORT void bulkUpdateDBRoot(const std::vector<BulkUpdateDBRootArg>&);
/** @brief Get HWM information about last segment file for each DBRoot
* assigned to a specific PM.
*
* Vector will contain an entry for each DBRoot. If no "available" extents
* are found for a DBRoot, then totalBlocks will be 0 (and hwmExtentIndex
* will be -1) for that DBRoot.
* @param OID The oid of interest.
* @param pmNumber The PM number of interest.
* @param emDbRootHwmInfos The vector of DbRoot/HWM related objects.
*/
EXPORT void getDbRootHWMInfo(int OID, uint16_t pmNumber, EmDbRootHWMInfo_v& emDbRootHwmInfos);
/** @brief Get the status (AVAILABLE, OUTOFSERVICE, etc) for the
* segment file represented by the specified OID, part# and seg#.
*
* Unlike many of the other DBRM functions, this function does
* not throw an exception if no extent is found; the "found"
* flag indicates whether an extent was found or not.
*
* @param oid (in) The OID of interest
* @param partitionNum (in) The partition number of interest
* @param segmentNum (in) The segment number of interest
* @param bFound (out) Indicates if extent was found or not
* @param status (out) The state of the extents in the specified
* segment file.
*/
EXPORT void getExtentState(int OID, uint32_t partitionNum, uint16_t segmentNum, bool& bFound, int& status);
/** @brief Gets the extents of a given OID
*
* Gets the extents of a given OID. The returned entries will
* be NULL-terminated and will have to be destroyed individually
* using delete.
* @note Untested
* @param OID (in) The OID to get the extents for.
* @param entries (out) A snapshot of the OID's Extent Map entries
* sorted by starting LBID; note that The Real Entries can change at
* any time.
* @param sorted (in) indicates if output is to be sorted
* @param notFoundErr (in) indicates if no extents is considered an err
* @param incOutOfService (in) include/exclude out of service extents
*/
EXPORT void getExtents(int OID, std::vector<struct EMEntry>& entries, bool sorted = true,
bool notFoundErr = true, bool incOutOfService = false);
/** @brief Gets the extents of a given OID under specified dbroot
*
* Gets the extents of a given OID under specified dbroot. The returned entries will
* be NULL-terminated and will have to be destroyed individually
* using delete.
* @param OID (in) The OID to get the extents for.
* @param entries (out) A snapshot of the OID's Extent Map entries for the dbroot
* @param dbroot (in) the specified dbroot
*/
EXPORT void getExtents_dbroot(int OID, std::vector<struct EMEntry>& entries, const uint16_t dbroot);
/** @brief Gets the number of extents for the specified OID and DBRoot
*
* @param OID (in) The OID of interest
* @param dbroot (in) The DBRoot of interest
* @param incOutOfService (in) include/exclude out of service extents
* @param numExtents (out) number of extents found for OID and dbroot
* @return 0 on success, non-0 on error (see brmtypes.h)
*/
EXPORT void getExtentCount_dbroot(int OID, uint16_t dbroot, bool incOutOfService, uint64_t& numExtents);
/** @brief Gets the size of an extent in rows
*
* Gets the size of an extent in rows.
* @return The number of rows in an extent.
*/
EXPORT unsigned getExtentSize(); // dmc-consider deprecating
EXPORT unsigned getExtentRows();
/** @brief Gets the DBRoot for the specified system catalog OID
*
* Function should only be called for System Catalog OIDs, as it assumes
* the OID is fully contained on a single DBRoot, returning the first
* DBRoot found. This only makes sence for a System Catalog
* OID, because all other column OIDs can span multiple DBRoots.
*
* @param oid The system catalog OID
* @param dbRoot (out) the DBRoot holding the system catalog OID
*/
EXPORT void getSysCatDBRoot(OID_t oid, uint16_t& dbRoot);
/** @brief Delete a Partition for the specified OID(s).
*
* @param oids (in) the OIDs of interest.
* @param partitionNums (in) the set of partitions to be deleted.
*/
EXPORT void deletePartition(const std::set<OID_t>& oids, const std::set<LogicalPartition>& partitionNums,
std::string& emsg);
/** @brief Mark a Partition for the specified OID(s) as out of service.
*
* @param oids (in) the OIDs of interest.
* @param partitionNums (in) the set of partitions to be marked out of service.
*/
EXPORT void markPartitionForDeletion(const std::set<OID_t>& oids,
const std::set<LogicalPartition>& partitionNums, std::string& emsg);
/** @brief Mark all Partition for the specified OID(s) as out of service.
*
* @param oids (in) the OIDs of interest.
*/
EXPORT void markAllPartitionForDeletion(const std::set<OID_t>& oids);
/** @brief Restore a Partition for the specified OID(s).
*
* @param oids (in) the OIDs of interest.
* @param partitionNums (in) the set of partitions to be restored.
*/
EXPORT void restorePartition(const std::set<OID_t>& oids, const std::set<LogicalPartition>& partitionNums,
std::string& emsg);
/** @brief Get the list of out-of-service partitions for a given OID
*
* @param OID (in) the OID of interest.
* @param partitionNums (out) the out-of-service partitions for the oid.
* partitionNums will be in sorted order.
*/
EXPORT void getOutOfServicePartitions(OID_t oid, std::set<LogicalPartition>& partitionNums);
/** @brief Delete all extent map rows for the specified dbroot
*
* @param dbroot (in) the dbroot
*/
EXPORT void deleteDBRoot(uint16_t dbroot);
/** @brief Is the specified DBRoot empty with no extents.
* Throws exception if extentmap shared memory is not loaded.
*
* @param dbroot DBRoot of interest
*/
EXPORT bool isDBRootEmpty(uint16_t dbroot);
/** @brief Performs internal consistency checks (for testing only)
*
* Performs internal consistency checks (for testing only).
* @note It's incomplete
* @return 0 if all tests pass, -1 (or throws logic_error) if not.
*/
EXPORT int checkConsistency();
EXPORT void setReadOnly();
EXPORT void undoChanges() override;
EXPORT void confirmChanges() override;
EXPORT int markInvalid(const LBID_t lbid, const execplan::CalpontSystemCatalog::ColDataType colDataType);
EXPORT int markInvalid(const std::vector<LBID_t>& lbids,
const std::vector<execplan::CalpontSystemCatalog::ColDataType>& colDataTypes);
EXPORT int setMaxMin(const LBID_t lbidRange, const int64_t max, const int64_t min, const int32_t seqNum,
bool firstNode);
// @bug 1970. Added setExtentsMaxMin function below.
/** @brief Updates the extents in the passed map of CPMaxMin objects.
* @param cpMap - The key must be the first LBID in the range.
* The values are a CPMaxMin struct with the
* min, max, and sequence.
* @param firstNode - if true, logs a debugging msg when CP data is updated
* @return 0 if all tests pass, -1 (or throws logic_error) if not.
*/
EXPORT void setExtentsMaxMin(const CPMaxMinMap_t& cpMap, bool firstNode, bool useLock = true);
/** @brief Merges the CP info for the extents contained in cpMap.
* @param cpMap - The key must be the starting LBID in the range.
* @return 0 if all tests pass, -1 (or throws logic_error) if not.
*/
void mergeExtentsMaxMin(CPMaxMinMergeMap_t& cpMap, bool useLock = true);
template <typename T>
EXPORT int getMaxMin(const LBID_t lbidRange, T& max, T& min, int32_t& seqNum);
EXPORT void getCPMaxMin(const LBID_t lbidRange,
CPMaxMin& cpMaxMin); /** @brief Get whole record for untyped use. */
inline bool empty()
{
if (fEMRBTreeShminfo == nullptr)
{
grabEMEntryTable(BRM::ExtentMap::READ);
releaseEMEntryTable(BRM::ExtentMap::READ);
}
// Initial size.
return (fEMRBTreeShminfo->currentSize == EM_RB_TREE_EMPTY_SIZE);
}
EXPORT std::vector<InlineLBIDRange> getFreeListEntries();
EXPORT void dumpTo(std::ostream& os);
EXPORT const std::atomic<bool>* getEMLockStatus();
EXPORT const std::atomic<bool>* getEMFLLockStatus();
EXPORT const std::atomic<bool>* getEMIndexLockStatus();
size_t EMIndexShmemSize();
size_t EMIndexShmemFree();
#ifdef BRM_DEBUG
EXPORT void printEM() const;
EXPORT void printEM(const OID_t& oid) const;
EXPORT void printEM(const EMEntry& em) const;
EXPORT void printFL() const;
#endif
private:
enum class UndoRecordType
{
DEFAULT,
INSERT,
DELETE
};
static const constexpr size_t EM_INCREMENT_ROWS = 1000;
static const constexpr size_t EM_INITIAL_SIZE = EM_INCREMENT_ROWS * 10 * sizeof(EMEntry);
static const constexpr size_t EM_INCREMENT = EM_INCREMENT_ROWS * sizeof(EMEntry);
static const constexpr size_t EM_FREELIST_INITIAL_SIZE = 50 * sizeof(InlineLBIDRange);
static const constexpr size_t EM_FREELIST_INCREMENT = 50 * sizeof(InlineLBIDRange);
static const constexpr size_t EM_SAVE_NUM_PER_BATCH = 1000000;
// RBTree constants.
static const size_t EM_RB_TREE_NODE_SIZE = sizeof(EMEntry) + 8 * sizeof(uint64_t);
static const size_t EM_RB_TREE_EMPTY_SIZE = 1024;
static const size_t EM_RB_TREE_INITIAL_SIZE = 16 * 1024 * 1024;
static const size_t EM_RB_TREE_INCREMENT = 16 * 1024 * 1024;
ExtentMap(const ExtentMap& em);
ExtentMap& operator=(const ExtentMap& em);
ExtentMapRBTree* fExtentMapRBTree;
InlineLBIDRange* fFreeList;
key_t fCurrentEMShmkey;
key_t fCurrentFLShmkey;
MSTEntry* fEMRBTreeShminfo;
MSTEntry* fFLShminfo;
MSTEntry* fEMIndexShminfo;
const MasterSegmentTable fMST;
bool r_only;
typedef std::tr1::unordered_map<int, oam::DBRootConfigList*> PmDbRootMap_t;
PmDbRootMap_t fPmDbRootMap;
time_t fCacheTime; // timestamp associated with config cache
int numUndoRecords;
std::atomic<bool> flLocked{false};
std::atomic<bool> emLocked{false};
std::atomic<bool> emIndexLocked{false};
static boost::mutex mutex; // @bug5355 - made mutex static
static boost::mutex emIndexMutex;
boost::mutex fConfigCacheMutex; // protect access to Config Cache
enum OPS
{
NONE,
READ,
WRITE
};
OPS EMLock, FLLock;
LastIndEmptyIndEmptyInd _createExtentCommonSearch(const OID_t OID, const DBRootT dbRoot,
const PartitionNumberT partitionNum,
const SegmentT segmentNum);
void logAndSetEMIndexReadOnly(const std::string& funcName);
// Create extents.
LBID_t _createColumnExtent_DBroot(uint32_t size, int OID, uint32_t colWidth, uint16_t dbRoot,
execplan::CalpontSystemCatalog::ColDataType colDataType,
uint32_t& partitionNum, uint16_t& segmentNum, uint32_t& startBlockOffset);
LBID_t _createColumnExtentExactFile(uint32_t size, int OID, uint32_t colWidth, uint16_t dbRoot,
uint32_t partitionNum, uint16_t segmentNum,
execplan::CalpontSystemCatalog::ColDataType colDataType,
uint32_t& startBlockOffset);
LBID_t _createDictStoreExtent(uint32_t size, int OID, uint16_t dbRoot, uint32_t partitionNum,
uint16_t segmentNum);
// Delete extent.
ExtentMapRBTree::iterator deleteExtent(ExtentMapRBTree::iterator it, const bool clearEMIndex = true);
template <typename T>
bool isValidCPRange(const T& max, const T& min, execplan::CalpontSystemCatalog::ColDataType type) const;
LBID_t getLBIDsFromFreeList(uint32_t size);
void reserveLBIDRange(LBID_t start, uint8_t size); // used by load() to allocate pre-existing LBIDs
std::vector<EMEntry> getEmIdentsByLbids(const bi::vector<LBID_t>& lbids);
std::vector<ExtentMapRBTree::iterator> getEmIteratorsByLbids(const bi::vector<LBID_t>& lbids);
// Choose keys.
key_t chooseEMShmkey();
key_t chooseFLShmkey();
key_t chooseEMIndexShmkey();
key_t getInitialEMIndexShmkey() const;
// see the code for how keys are segmented
key_t chooseShmkey(const MSTEntry* masterTableEntry, const uint32_t keyRangeBase) const;
// Grab table.
void grabEMEntryTable(OPS op);
void grabFreeList(OPS op);
void grabEMIndex(OPS op);
// Release table.
void releaseEMEntryTable(OPS op);
void releaseFreeList(OPS op);
void releaseEMIndex(OPS op);
// Grow memory.
void growEMShmseg(size_t nrows = 0);
void growFLShmseg();
void growEMIndexShmseg(const size_t suggestedSize = 0);
void growIfNeededOnExtentCreate();
// Finish.
void finishChanges();
EXPORT unsigned getFilesPerColumnPartition();
unsigned getExtentsPerSegmentFile();
unsigned getDbRootCount();
void getPmDbRoots(int pm, std::vector<int>& dbRootList);
DBRootVec getAllDbRoots();
void checkReloadConfig();
ShmKeys fShmKeys;
void makeUndoRecordRBTree(UndoRecordType type, const EMEntry& emEntry);
void undoChangesRBTree();
void confirmChangesRBTree();
bool fDebug;
int _markInvalid(const LBID_t lbid, const execplan::CalpontSystemCatalog::ColDataType colDataType);
MSTEntry* _getTableLock(const OPS op, std::atomic<bool>& lockedState, const int table);
void _getTableLockUpgradeIfNeeded(const OPS op, std::atomic<bool>& lockedState, const int table);
void _getTableLockDowngradeIfNeeded(const OPS op, std::atomic<bool>& lockedState, const int table);
void _releaseTable(const OPS op, std::atomic<bool>& lockedState, const int table);
template <typename T>
void load(T* in);
template <typename T>
void loadVersion4or5(T* in, bool upgradeV4ToV5);
ExtentMapRBTree::iterator findByLBID(const LBID_t lbid);
using UndoRecordPair = std::pair<UndoRecordType, EMEntry>;
std::vector<UndoRecordPair> undoRecordsRBTree;
ExtentMapRBTreeImpl* fPExtMapRBTreeImpl;
FreeListImpl* fPFreeListImpl;
ExtentMapIndexImpl* fPExtMapIndexImpl_;
};
inline std::ostream& operator<<(std::ostream& os, ExtentMap& rhs)
{
rhs.dumpTo(os);
return os;
}
} // namespace BRM
#undef EXPORT
View Git Blame Copy Permalink