8.1.10 The rpd_tables Table

The rpd_tables table provides the system change number (SCN) and load pool type for tables loaded in MySQL HeatWave.

The rpd_tables table is read-only.

The rpd_tables table has these columns:

• ID

  A unique identifier for the table.

• SNAPSHOT_SCN

  The system change number (SCN) of the table snapshot. The SCN is an internal number that represents a point in time according to the system logical clock that the table snapshot was transactionally consistent with the source table.

• PERSISTED_SCN

  The SCN up to which changes are persisted.

• POOL_TYPE

  The load pool type of the table.

  • As of MySQL 9.5.0, the possible values are SNAPSHOT, TRANSACTIONAL, and VOLATILE.

  • As of MySQL 9.2.1, the possible values are SNAPSHOT and TRANSACTIONAL.

  • In previous versions of MySQL, the possible values are RAPID_LOAD_POOL_SNAPSHOT and RAPID_LOAD_POOL_TRANSACTIONAL.

• DATA_PLACEMENT_TYPE

  The data placement type.

• NROWS

  The number of rows that are loaded for the table. The value is set initially when the table is loaded, and updated as changes are propagated.

• LOAD_STATUS

  The load status of the table. Statuses include:

  • NOLOAD_RPDGSTABSTATE

    The table is not yet loaded.

  • LOADING_RPDGSTABSTATE

    The table is being loaded.

  • AVAIL_RPDGSTABSTATE

    The table is loaded and available for queries.

  • UNLOADING_RPDGSTABSTATE

    The table is being unloaded.

  • INRECOVERY_RPDGSTABSTATE

    The table is being recovered. After completion of the recovery operation, the table is placed back in the UNAVAIL_RPDGSTABSTATE state if there are pending recoveries.

  • STALE_RPDGSTABSTATE

    A failure during change propagation, and the table has become stale. See Section 4.2.7, “About Change Propagation”

    Note

    When TRUNCATE TABLE operation is performed, the table is marked for a full reload. However, the table is not reloaded immediately. The system reloads the table only after all currently undergoing incremental or scheduled reloads. hence, the table may appear as stale in the system (e.g., in RPDGSTABSTATE) until it is reloaded completely.

  • UNAVAIL_RPDGSTABSTATE

    The table is unavailable.

  • RECOVERYFAILED_RPDGSTABSTATE

    This status is available as of MySQL 9.2.2. A Lakehouse table is marked as RECOVERYFAILED_RPDGSTABSTATE when an attempt to recover it fails. Tables in this state can be unloaded. Therefore, before using tables in this state, you need to manually reload them using SECONDARY_UNLOAD followed by SECONDARY_LOAD. When a MySQL HeatWave cluster restarts, there is no attempt to recover tables in RECOVERYFAILED_RPDGSTABSTATE. However, when a DB System restarts, an attempt is made to recover tables in this state.

• LOAD_PROGRESS

  The load progress of the table expressed as a percentage value.

  For Lakehouse, the following values are returned:

  • 10%: the initialization phase is complete.

  • 10-70%: the transformation to native MySQL HeatWave format is in progress.

  • 70% - 80%: the transformation to native MySQL HeatWave format is complete and the aggregation phase is in progress.

  • 80-99%: the recovery phase is in progress.

  • 100%: data load is complete.

• SIZE_BYTES

  The amount of data loaded for the table, in bytes.

• TRANSFORMATION_BYTES:

  The total size of raw Lakehouse data transformed, in bytes.

• QUERY_COUNT

  The number of queries that referenced the table.

• LAST_QUERIED

  The timestamp of the last query that referenced the table.

• LOAD_START_TIMESTAMP

  The load start timestamp for the table.

• LOAD_END_TIMESTAMP

  The load completion timestamp for the table.

• RECOVERY_SOURCE

  Indicates the source of the last successful recovery for a table. The values are MySQL, that is InnoDB, or ObjectStorage.

• RECOVERY_START_TIMESTAMP

  The timestamp when the latest successful recovery started.

• RECOVERY_END_TIMESTAMP

  The timestamp when the latest successful recovery ended.

• LOAD_TYPE

  Specifies whether the table is automatically loaded by MySQL HeatWave (SELF) or loaded by the user (USER).

• LOGICAL_PARTS_LOADED_AT_SCN

  For partitioned tables, contains an array of JSON objects, each corresponding to a loaded partition. For each partition, the following values are stored:

  • id: the id of the partition in InnoDB.

  • name: the name of the partition.

  • load_scn: the system change number (SCN) at which the partition was loaded.

  • load_type: specifies whether the table is automatically loaded by MySQL HeatWave, or loaded by the user.

• AUTO_ZMP_COLUMNS

  Contains a list of IDs that correspond to the columns on which zone maps are automatically built.

• ACE_MODEL

  Available as of MySQL 9.3.0. Indicates whether an Advanced Cardinality Estimation (ACE) statistics model is currently associated with the given table in MySQL HeatWave.

• STALE_REASON

  Available as of MySQL 9.7.0. Identifies the reason that a table became stale.

  It includes the following reasons:

  • OK

    The table is not stale.

  • ERROR_CLUSTER_OOM

    The MySQL HeatWave Cluster ran out of memory.

  • ERROR_PKEY_NOT_FOUND

    The old version of an updated or deleted row was not found.

  • ERROR_UU_OVERLAP

    Update units with overlapping primary keys were found.

  • ERROR_INVALID_CP_PACKET

    An invalid change propagation packet.

  • ERROR_UU_TOO_LARGE

    An update unit that was larger than expected was found.

  • DECOMPRESSION_TIMEOUT

    The decompression of data took too long and timed out.

  • ERROR_CLUSTER_OTHER

    A non-specific error with the MySQL HeatWave Cluster.

  • TABLE_TYPE_MISMATCH

    An unexpected type of table was found (InnoDB versus Lakehouse).

  • RPD_CP_RESET

    Change propagation encountered a systemic error and had to reset.

  • DROP_PARTITION_FAILED

    An operation to drop a partition failed.

  • ADD_PARTITION_FAILED

    An operation to add a partition failed.

  • ALTER_PARTITION_FAILED

    An operation to alter a partition failed.

  • PARTITION_UNLOAD_FAILED

    An operation to unload a partition failed.

  • PARTITION_LOAD_FAILED

    An operation to load a partition failed.

  • RELOAD_REQUIRED

    The change propagation operation requires a reload.

  • RPD_PARSING_OOM

    Binlog parsing ran out of memory.

  • RPD_PARSER_ERROR

    Binlog parsing encountered a parsing error.

  • UNIDENTIFIED_ERROR

    An unexpected error occurred during change propagation.

  If tables become stale, you must unload them and load them again. To learn more, see the following:

  • Unload Data from MySQL HeatWave Cluster

  • Load Data Using Auto Parallel Load

  • Load Data Manually