All Products
Search

This Product

    PolarDB:Default values of cluster parameters

    Document Center

    PolarDB:Default values of cluster parameters

    Last Updated:Nov 23, 2024

    This topic describes the default values of cluster parameters.

    Parameters

    The default values of PolarDB for PostgreSQL parameters are a result of considering performance, availability, and reliability. They may be different from the default values of the community version. In particular, the default values of the following parameters are based on the principle of optimal performance. You can modify these parameters in the PolarDB console.

    Parameter

    Default value

    Description

    synchronous_commit

    off

    Specifies whether a transaction needs to wait for WAL records to be written to the disk before it returns a success message to the client. Default value: off. Valid values:

    • off: A transaction does not wait for WAL records to be written to the disk before it returns a success message to the client.

      Note

      Latency exists between returning success to the client and true transaction persistence. The maximum latency value is three times the value of the wal_writer_delay parameter. The default value of the wal_writer_delay parameter is 10ms. When this parameter is set to off, the transaction commit efficiency can be improved.

    • on: A transaction waits for WAL records to be written to the disk before it returns a success message to the client.

      Note

      If persistence on disk is required after a transaction is committed, you can set this parameter to on in the console.

    • remote_apply: A transaction waits for WAL records to be written to the disk before it returns a success message to the client. After you set this value for the synchronous_commit parameter, commits wait until replies from the current synchronous standby(s) indicate they have received the commit record of the transaction and applied it, so that it has become visible to queries on the standby(s), and also written to durable storage on the standbys. This will cause much larger commit delays than previous settings because it waits for WAL replay.

    • remote_write: A transaction waits for WAL records to be written to the file system before it returns a success message to the client.

    • local: A transaction waits for WAL records to be written to the disk, but a success message is not required to be sent to the client.

    Note

    You can execute the following statement to view the settings of this parameter for the current cluster:

    SHOW synchronous_commit;

    wal_level

    replica

    Specifies what types of data are written to WAL. Default value: replica. Valid values: Default value: replica. Valid values:

    • replica: The data required for logical replication is not written to WAL. This reduces the volume of data written to WAL and improves data write performance.

    • logical: The data required for logical replication is written to WAL. This increases the volume of data written to WAL.

    Warning

    The cluster restarts after you modify this parameter. Proceed with caution.

    Note

    You can execute the following statement to view the settings of this parameter for the current cluster:

    SHOW wal_level;

    log_statement

    ddl

    Specifies the level of audit logs. Default value: ddl, which indicates that only logs for DDL statements are audited. Displaying less logs can improve database performance.

    Note

    • You can enable the audit log feature on the SQL Explorer page of the console. Then, logs for all SQL statements are audited. For more information, see ️SQL Explorer.

    • You can execute the following statement to view the settings of this parameter for the current cluster:

      SHOW log_statement;

    Default values of common cluster parameters

    Parameter

    Default value

    authentication_timeout

    60

    autovacuum

    on

    autovacuum_analyze_scale_factor

    0.05

    autovacuum_analyze_threshold

    50

    autovacuum_freeze_max_age

    500000000

    autovacuum_max_workers

    5

    autovacuum_multixact_freeze_max_age

    700000000

    autovacuum_naptime

    30

    autovacuum_vacuum_cost_delay

    0

    autovacuum_vacuum_cost_limit

    10000

    autovacuum_vacuum_scale_factor

    0.02

    autovacuum_vacuum_threshold

    50

    backend_flush_after

    0

    bgwriter_delay

    10

    checkpoint_timeout

    30

    constraint_exclusion

    partition

    cpu_index_tuple_cost

    0.005

    cpu_operator_cost

    0.0025

    cpu_tuple_cost

    0.01

    datestyle

    'ISO,YMD'

    deadlock_timeout

    1000

    default_with_oids

    off

    dynamic_shared_memory_type

    posix

    enable_partition_pruning

    on

    extra_float_digits

    0

    fsync

    on

    full_page_writes

    off

    gin_pending_list_limit

    4096

    hot_standby

    on

    hot_standby_feedback

    on

    huge_pages

    on

    idle_in_transaction_session_timeout

    3600000

    jit

    off

    lock_timeout

    0

    logging_collector

    on

    log_connections

    off

    log_disconnections

    off

    log_error_verbosity

    default

    log_min_duration_statement

    5000

    log_statement

    ddl

    max_files_per_process

    10000

    max_locks_per_transaction

    64

    max_prepared_transactions

    800

    max_replication_slots

    64

    max_worker_processes

    256

    seq_page_cost

    1

    ssl

    off

    statement_timeout

    0

    synchronous_commit

    off

    temp_file_limit

    100 GB

    timezone

    'UTC'

    wal_buffers

    16 MB

    wal_level

    replica

    wal_writer_delay

    10