OceanBaseデータベース

SQL - V4.4.2

PDF をダウンロード

GV$OB_MEMORY|V5.0.1|OceanBase Database|docs|OceanBase Documentations

このページには現在中国語/日本語版がなく、英語版を表示しています。

GV$OB_MEMORY

最終更新:2026-09-04 11:28:25

Note

This view was renamed from GV$MEMORY to GV$OB_MEMORY starting with V4.0.0.

Purpose

This view displays the memory statistics of all tenants on all nodes.

Columns

Column	Type	Nullable	Description
TENANT_ID	bigint(20)	NO	The tenant ID.
SVR_IP	varchar(46)	NO	The IP address of the server.
SVR_PORT	bigint(20)	NO	The port number of the server.
CTX_NAME	varchar(256)	NO	The name of the CTX to which the memory belongs.
MOD_NAME	varchar(256)	NO	The name of the Mod to which the memory belongs.
COUNT	decimal(20,0)	NO	The difference between the allocated and released memory, that is, the number of memory units currently used by the Mod.
HOLD	decimal(20,0)	NO	The size of memory currently occupied by the Mod, in bytes.
USED	decimal(20,0)	NO	The size of memory currently used by the Mod, in bytes.

CTX_NAME values

The following table describes the names and meanings of common memory CTXs.

CTX_NAME	Description
DEFAULT_CTX_ID	The ID of the memory context for unallocated memory.
MEMSTORE_CTX_ID	The ID of the memory context for user MemTable data.
TRANS_CTX_MGR_ID	The ID of the transaction context.
PLAN_CACHE_CTX_ID	The ID of the execution plan cache context.
GLIBC	The memory allocated by using the Glibc Malloc interface.
CO_STACK	The memory for the thread execution stack.
KVSTORE_CACHE_ID	The ID of the dynamic scalable cache.
META_OBJ_CTX_ID	The memory for storing metadata, such as macroblock metadata.
TX_CALLBACK_CTX_ID	The ID of the transaction callback context. This context is used to record some modifications during transaction execution. When a transaction is committed or rolled back, the corresponding data can be found and operated on.
LOB_CTX_ID	The Allocator ID in the ObLobManager.
PS_CACHE_CTX_ID	The cache ID for the `PREPARE stmt_name FROM preparable_stmt` statement.
TX_DATA_TABLE	The transaction data table. It stores the status of resolved transactions and is used for dumping queries in scenarios such as uncommitted transactions and FAST COMMIT.
MDS_DATA_ID	The memory ID of multi-source data
LIBEASY	The memory of the network framework (EASY).
LOGGER_CTX_ID	The ID of the program log memory context.
RPC_CTX_ID	The ID of the RPC framework memory context.
PKT_NIO	The memory of the network framework (pkt_nio).
SCHEMA_SERVICE	The memory for storing schema metadata, such as tables and databases.

Correspondence between CTX_NAME and MOD_NAME

The following table lists the correspondence between CTX_NAME and MOD_NAME.

CTX_NAME	MOD_NAME	TENANT
DEFAULT_CTX_ID	MysqlRequesReco: MySQL request record module LogAggreBuffer: log aggregation buffer module PartitLogServic: partition log service module SqlDtl: SQL detail module TransAudit: transaction audit module IoControl: I/O control module ResultSet: query result set module	ALL
DEFAULT_CTX_ID	SchemaSysCache: schema system cache module SeArray: sorted array module SstaMicrBlocAll: statistics, microblock, and other modules CsSstableReader: compressed table reader TmpBlockManager: temporary block manager TenantSchemMgr: tenant schema manager DecoderCtx: decoder context SqlSqcHandler: SQL semantic processor	500
SQL_EXEC_CTX_ID	CostBasedRewrit: cost-based rewrite optimizer SqlExecutor: SQL executor SqlPsCache: SQL parameter cache module OB_SQL_AGGR_FUNC_ROW: SQL aggregate function row object	ALL
WORK_AREA	SqlAggrFuncRow: SQL aggregate function row object WindowAggProc: window aggregate processing module SqLWindoRowStor: window row storage module SqLChunkRowStor: chunk row storage module SqLTempTabLeRow: temporary table row module ConnectByPump: data transfer module for connect operations SqlSortRow: sorted row module SortOpRows: sorted row object collection SqLHashDisRowSt: hash distributed row storage module SqLHashDist: hash distribution module HashNodeGrouRow: hash node grouping row module HashDistinctOp: hash deduplication operation module ArenaHashJoin: hash join memory space management module HtOpAlloc: hash table operation allocation module
PLAN_CACHE_CTX_ID	OB_SQL_PHY_PLAN: query plan module OB_SQL_PLAN_CACHE: query plan cache module OB_SQL_PHY_PL_OBJ: query plan object module PRE_CALC_EXPR: precomputed expression module
TRANS_CTX_MGR_ID	PartTranCtxMgr: partitioned transaction context TenantCtxAlloca: tenant allocation context
TX_CALLBACK_CTX_ID	MemtableCallbac: memory table callback module
REPLAY_STATUS_CTX_ID	ReplayStatus: replay status module
CO_STACK	CO_STACK: coroutine stack module	500
STORAGE_LONG_TERM_META_CTX_ID	Partition: partition object PartitioStore: partition storage module PartitioStorage: partition storage
STORAGE_SHORT_TERM_META_CTX_ID	Sstable: SSTable table SstableMeta: meta tenant SSTable table MemtableObject: memory table object
ALL	OB_COMMON_ARRAY: OceanBase common array	ALL

Impact of common SQL operations on memory CTX

In OceanBase Database, each SQL execution or operation corresponds to a specific CTX. The size of the data and the complexity of the operation can affect the memory usage of the CTX. The following are some common SQL operations that may lead to increased memory usage in certain CTXs:

Sorting large amounts of data (ORDER BY, GROUP BY, etc.): Sorting operations require loading data into memory for sorting, which may increase the memory usage in PLAN_CACHE_CTX_ID, MEMSTORE_CTX_ID, and DEFAULT_CTX_ID.
Aggregating large amounts of data (SUM, AVG, MAX, MIN, etc.): Aggregation operations require loading data into memory for calculations, which may increase the memory usage in PLAN_CACHE_CTX_ID, MEMSTORE_CTX_ID, and DEFAULT_CTX_ID.
Joining large amounts of data (JOIN): Join operations require loading data into memory for matching, which may increase the memory usage in PLAN_CACHE_CTX_ID, MEMSTORE_CTX_ID, and DEFAULT_CTX_ID.
Querying large amounts of data (SELECT): Query operations require loading data into memory for filtering and sorting, which may increase the memory usage in PLAN_CACHE_CTX_ID, MEMSTORE_CTX_ID, and DEFAULT_CTX_ID.
Writing large amounts of data (INSERT, UPDATE, DELETE, etc.): Write operations require managing and processing transactions, which may increase the memory usage in TRANS_CTX_MGR_ID, MEMSTORE_CTX_ID, and DEFAULT_CTX_ID.
Logging large amounts of data (REDO, UNDO, BINLOG, etc.): Logging operations require loading data into memory for operations, which may increase the memory usage in TRANS_CTX_MGR_ID, DEFAULT_CTX_ID, LOB_CTX_ID, and TX_DATA_TABLE.

次のトピック