V$OB_MEMORY|V4.3.0| docs|Distributed Database

V$OB_MEMORY

Last Updated：2024-04-19 08:42:49 Updated

Note

The view name is changed from V$MEMORY to V$OB_MEMORY since OceanBase Database V4.0.0.

Purpose

The V$OB_MEMORY view displays the tenant-level memory statistics.

Columns

Column	Type	Nullable?	Description
TENANT_ID	bigint(20)	NO	The ID of the tenant.
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 context to which the memory belongs.
MOD_NAME	varchar(256)	NO	The name of the module to which the memory belongs.
COUNT	decimal(20,0)	NO	The difference between memory spaces that are allocated and released. That is, the number of memory units used by the module.
HOLD	decimal(20,0)	NO	The memory size currently occupied by the module.
USED	decimal(20,0)	NO	The size in bytes of memory being used by the module.

Values of the CTX_NAME column

The following table lists the names and meanings of general memory contexts.

CTX_NAME	Description
DEFAULT_CTX_ID	The context ID of unallocated memory.
MEMSTORE_CTX_ID	The MemTable data context ID.
TRANS_CTX_MGR_ID	The transaction context ID.
PLAN_CACHE_CTX_ID	The context ID of the plan cache.
GLIBC	The memory applied for by using the Glibc Malloc API.
CO_STACK	The memory for the execution stack of threads.
KVSTORE_CACHE_ID	The ID of the auto-scaling cache.
META_OBJ_CTX_ID	The memory that stores metadata, such as metadata of macroblocks.
TX_CALLBACK_CTX_ID	The ID of the transaction callback context. This context records some modifications made during transaction execution so that the required data can be located when a transaction is committed or rolled back.
LOB_CTX_ID	The allocator ID in ObLobManager.
PS_CACHE_CTX_ID	The cache ID of the `PREPARE stmt_name FROM preparable_stmt` statement.
TX_DATA_TABLE	The transaction data table. This table stores the status of committed or rolled back transactions and is used in scenarios such as minor compaction of status information of uncommitted transactions and fast commits.
MDS_DATA_ID	The memory ID of multi-source data.
LIBEASY	The network framework memory (EASY).
LOGGER_CTX_ID	The context ID of the memory for application logs.
RPC_CTX_ID	The context ID of the memory for the RPC framework.
PKT_NIO	The network framework memory (pkt_nio).
SCHEMA_SERVICE	The memory for schema metadata, such as tables and databases.

Mappings between CTX_NAME and MOD_NAME

The following table lists the mappings between CTX_NAME and MOD_NAME.

CTX_NAME	MOD_NAME	TENANT
DEFAULT_CTX_ID	MysqlRequesReco: the module for recording MySQL requests. LogAggreBuffer: the buffer module for log aggregation. PartitLogServic: the partition log service module. SqlDtl: the SQL details module. TransAudit: the transaction audit module. IoControl: the input/output control module. ResultSet: the module where result sets are located.	ALL
DEFAULT_CTX_ID	SchemaSysCache: the system-level cache module that caches mode information. SeArray: the sorting array module. SstaMicrBlocAll: the module for statistics and microblocks. CsSstableReader: the compressed table reader. TmpBlockManager: the temporary block manager. TenantSchemMgr: the tenant mode manager. DecoderCtx: the decoder context. SqlSqcHandler: the SQL semantics handler.	500
SQL_EXEC_CTX_ID	CostBasedRewrit: the cost-based rewrite optimizer. SqlExecutor: the SQL executor. SqlPsCache: the SQL parameter cache module. OB_SQL_AGGR_FUNC_ROW: the row object for SQL aggregate functions.	ALL
WORK_AREA	SqlAggrFuncRow: the row object for SQL aggregate functions. WindowAggProc: the module for window aggregation processing. SqLWindoRowStor: the module related to rows in the storage window. SqLChunkRowStor: the module related to rows in storage chunks. SqLTempTabLeRow: the module related to rows in temporary tables. ConnectByPump: the data transmission module for connections. SqlSortRow: the row sorting module. SortOpRows: a collection of sorted row objects. SqLHashDisRowSt: the module for storing hash-distributed rows. SqLHashDist: the hash distribution module. HashNodeGrouRow: the module related to grouped rows on hash nodes. HashDistinctOp: the module for deduplicating hashed data. ArenaHashJoin: the memory space management module for HASH JOIN operations. HtOpAlloc: the module for the allocation of hash table operations.
PLAN_CACHE_CTX_ID	OB_SQL_PHY_PLAN: the execution plan module. OB_SQL_PLAN_CACHE: the module that caches execution plans. OB_SQL_PHY_PL_OBJ: the execution plan object module. PRE_CALC_EXPR: the module related to precomputation expressions.
TRANS_CTX_MGR_ID	PartTranCtxMgr: the partition transaction context. TenantCtxAlloca: the tenant allocation context.
TX_CALLBACK_CTX_ID	MemtableCallbac: the module related to callbacks of memory tables.
REPLAY_STATUS_CTX_ID	ReplayStatus: the module related to the replay status.
CO_STACK	CO_STACK: the module related to the coroutine stack.	500
STORAGE_LONG_TERM_META_CTX_ID	Partition: a partition object. PartitioStore: the partition store module. PartitioStorage: partition storage.
STORAGE_SHORT_TERM_META_CTX_ID	Sstable: the SSTable. SstableMeta: the SSTable of the meta tenant. MemtableObject: the memory table object.
ALL	OB_COMMON_ARRAY: the common array in OceanBase Database.	ALL

Impact of general SQL operations on the memory context

In OceanBase Database, each SQL execution or operation corresponds to a context. The memory usage varies based on the memory context used, depending on the data volume and operation complexity. The following general SQL executions or operations may cause the increase in memory usage of some contexts.

Sorting of a large amount of data by using the ORDER BY or GROUP BY clause: A sorting operation needs to load the data to the memory for sorting. This may cause an increase in the memory usage of PLAN_CACHE_CTX_ID, MEMSTORE_CTX_ID, and DEFAULT_CTX_ID.
Aggregation of a large amount of data by using the SUM, AVG, MAX, or MIN function: An aggregation operation needs to load the data to the memory for computing. This may cause an increase in the memory usage of PLAN_CACHE_CTX_ID, MEMSTORE_CTX_ID, and DEFAULT_CTX_ID.
Join of a large amount of data by using JOIN operations: A join operation needs to load the data to the memory for matching. This may cause an increase in the memory usage of PLAN_CACHE_CTX_ID, MEMSTORE_CTX_ID, and DEFAULT_CTX_ID.
Query of a large amount of data by using the SELECT statement: A query operation needs to load the data to the memory for filtering and sorting. This may cause an increase in the memory usage of PLAN_CACHE_CTX_ID, MEMSTORE_CTX_ID, and DEFAULT_CTX_ID.
Write of a large amount of data by using the INSERT, UPDATE, or DELETE statement: A write operation requires transaction management and processing. This may cause an increase in the memory usage of TRANS_CTX_MGR_ID, MEMSTORE_CTX_ID, and DEFAULT_CTX_ID.
Logging of a large amount of data into the REDO log, UNDO log, or binlog: A logging operation needs to load the data to the memory. This may cause an increase in the memory usage of TRANS_CTX_MGR_ID, DEFAULT_CTX_ID, LOB_CTX_ID, and TX_DATA_TABLE.