Managing Parameters

Last updated:2025-10-23 19:19:42

Managing Parameters

Last updated: 2025-10-23 19:19:42

This document describes the parameters that can be modified for read-only analysis engine instances.
Prerequisites
A read-only analysis engine instance is prepared. For details, see Enabling Read-Only Analysis Engine.
The instance is running.
Managing Parameters
Parameters for read-only analysis engine instances can be managed, including parameter modification, parameter template application, and parameter change history viewing. The operation methods are generally the same as those for read-only instances with the TXSQL engine. For details, see Setting Instance Parameters.
Parameter List
The following table lists the modifiable parameters for read-only analysis engine instances and recommendations for modifying parameters.
Storage Engine Parameters
Parameter Name
Default Value
Restart Required
Parameter Modification Recommendations
block_cache_capacity_mb
512
Yes
This parameter controls the cache size of the read-only analysis engine in the unit of MB. A larger cache space can cache more table data, improving execution efficiency after SQL statements are executed multiple times by avoiding repeated disk data scans.
max_threads
16
No
This parameter controls the maximum number of threads that can be used for SQL execution on a read-only analysis engine. By default, it is set to match the maximum number of CPUs of the instance. It is recommended to set the value to half of the maximum number of CPUs.
max_block_size
16
No
This parameter controls the size of a single data block during vectorized execution in the unit of MB. A larger value may result in excessive scan costs in scenarios with fine-grained queries. However, a smaller value may result in numerous quick scans, which can degrade performance. Tuning of this parameter should focus on the query scenarios. If most queries involve large-scale data scans, a larger value is appropriate. If most queries are simple point queries or involve small result set queries, a smaller value is appropriate.
Row and Column Data Loading Parameters
Parameter Name
Default Value
Restart Required
Parameter Modification Recommendations
replication_sink_parallelism
16
No
Modifying this parameter increases the concurrency level during incremental data loading, which can accelerate data loading. However, it also increases the load on the read-write instance. It is recommended to adjust this parameter only when the read-write instance has sufficient available resources.
replication_dumper_table_parallel
4
No
This parameter controls the number of tables on which full data loading can be performed. When the read-write instance has sufficient resources, increasing the parameter value can accelerate full data loading. It is recommended to adjust this parameter only when the read-write instance has sufficient available resources.
replication_dumper_insert_parallel
4
No
This parameter controls the concurrency level for writing data to a read-only analysis engine during full data loading. When a read-only analysis engine has sufficient resources, a higher concurrency level can accelerate full data loading. It is recommended to adjust this parameter only when the read-only analysis engine has sufficient available resources.
replication_dumper_select_parallel
16
No
This parameter controls the concurrency level for reading table data during full data loading. When the read-write instance has sufficient resources, increasing the concurrency level can accelerate full data loading. It is recommended to adjust this parameter only when the read-write instance has sufficient available resources.
replication_dumper_chunk_size
10000
No
This parameter controls the size of each chunk read from the read-write instance during full data loading. A larger chunk size indicates faster full data loading and higher resource consumption. It is recommended to adjust this parameter only when the read-write instance has sufficient available resources.
Computing Engine Parameters
Parameter Name
Default Value
Whether to Restart
Parameter Modification Suggestion
group_concat_max_len
1024
No
Maximum input parameter length in the unit of bytes allowed by the GROUP_CONCAT() function.
libra_query_log_max_len
4096
No
This parameter controls the maximum length of SQL records in slow log management in the unit of KB.
libra_max_delay_allow_query
86400
No
This parameter controls the maximum delay allowed for the objects involved in executing SQL statements in the analysis engine in the unit of seconds. If this time is exceeded, an error will be reported indicating that the delay value is too large during SQL execution in the analysis engine.
libra_etl_concurrency
1
No
Controls the number of concurrent threads for data writeback to the read-write instance. This parameter takes effect only when libra_concurrent_etl is set to ON.
libra_concurrent_etl
OFF
No
Controls whether to enable concurrent writeback to the read-write instance.
libra_etl_to_rw
OFF
No
Controls whether to enable the ETL writeback acceleration feature.
libra_enable_snapshot_read
ON
No
Controls whether to enable the snapshot read feature for the analysis engine.
libra_max_mpp_query_num
0
No
This parameter controls the number of SQL statements that the analysis engine is allowed to execute concurrently. The default value is 0, which indicates no limitations. It is generally enabled to avoid excessive concurrency that can consume all database resources. Once enabled, any subsequent SQL statements that exceed the allowed concurrent execution limit will be rejected for execution.
max_execution_time
0
No
This parameter controls the maximum execution time allowed for SQL statements in the unit of milliseconds. SQL statements exceeding this execution time will be automatically interrupted. The default value is 0, indicating that execution will not be interrupted.
libra_preserve_order_for_pagination
OFF
No
This parameter controls whether to enable the ordered pagination feature globally. For detailed feature introduction, see Ordered Pagination Feature.
libra_runtime_filter_type
MIN_MAX,BLOOM_FILTER,IN_FILTER
No
Indicates the settings of the Runtime filter types that can be assigned.
runtime_filter_max_in_num
1024
No
Maximum NDV value allowed when an IN-type Runtime Filter is generated.
libra_enable_runtime_filter
ON
No
Controls whether to enable or disable the Runtime Filter feature. For more details, see Runtime Filter User Manual.
sql_mode
Consistent with the read-write instance
No
This parameter controls SQL execution behavior in the analysis engine. Its behavior logic is consistent with that of MySQL. However, certain SQL_MODE values are either incompatible or configurable but not effective. For details, see the SQL_MODE section in Compatibility and Use Limits.
libra_enable_late_materialization
ON
No
Controls whether to enable or disable the lazy materialization feature. For more details, see Lazy Materialization.

Was this page helpful?

You can also Contact Sales or Submit a Ticket for help.

Yes

Feedback

Parameter Name	Default Value	Restart Required	Parameter Modification Recommendations
block_cache_capacity_mb	512	Yes	This parameter controls the cache size of the read-only analysis engine in the unit of MB. A larger cache space can cache more table data, improving execution efficiency after SQL statements are executed multiple times by avoiding repeated disk data scans.
max_threads	16	No	This parameter controls the maximum number of threads that can be used for SQL execution on a read-only analysis engine. By default, it is set to match the maximum number of CPUs of the instance. It is recommended to set the value to half of the maximum number of CPUs.
max_block_size	16	No	This parameter controls the size of a single data block during vectorized execution in the unit of MB. A larger value may result in excessive scan costs in scenarios with fine-grained queries. However, a smaller value may result in numerous quick scans, which can degrade performance. Tuning of this parameter should focus on the query scenarios. If most queries involve large-scale data scans, a larger value is appropriate. If most queries are simple point queries or involve small result set queries, a smaller value is appropriate.

Parameter Name	Default Value	Whether to Restart	Parameter Modification Suggestion
group_concat_max_len	1024	No	Maximum input parameter length in the unit of bytes allowed by the GROUP_CONCAT() function.
libra_query_log_max_len	4096	No	This parameter controls the maximum length of SQL records in slow log management in the unit of KB.
libra_max_delay_allow_query	86400	No	This parameter controls the maximum delay allowed for the objects involved in executing SQL statements in the analysis engine in the unit of seconds. If this time is exceeded, an error will be reported indicating that the delay value is too large during SQL execution in the analysis engine.
libra_etl_concurrency	1	No	Controls the number of concurrent threads for data writeback to the read-write instance. This parameter takes effect only when libra_concurrent_etl is set to ON.
libra_concurrent_etl	OFF	No	Controls whether to enable concurrent writeback to the read-write instance.
libra_etl_to_rw	OFF	No	Controls whether to enable the ETL writeback acceleration feature.
libra_enable_snapshot_read	ON	No	Controls whether to enable the snapshot read feature for the analysis engine.
libra_max_mpp_query_num	0	No	This parameter controls the number of SQL statements that the analysis engine is allowed to execute concurrently. The default value is 0, which indicates no limitations. It is generally enabled to avoid excessive concurrency that can consume all database resources. Once enabled, any subsequent SQL statements that exceed the allowed concurrent execution limit will be rejected for execution.
max_execution_time	0	No	This parameter controls the maximum execution time allowed for SQL statements in the unit of milliseconds. SQL statements exceeding this execution time will be automatically interrupted. The default value is 0, indicating that execution will not be interrupted.
libra_preserve_order_for_pagination	OFF	No	This parameter controls whether to enable the ordered pagination feature globally. For detailed feature introduction, see Ordered Pagination Feature.
libra_runtime_filter_type	MIN_MAX,BLOOM_FILTER,IN_FILTER	No	Indicates the settings of the Runtime filter types that can be assigned.
runtime_filter_max_in_num	1024	No	Maximum NDV value allowed when an IN-type Runtime Filter is generated.
libra_enable_runtime_filter	ON	No	Controls whether to enable or disable the Runtime Filter feature. For more details, see Runtime Filter User Manual.
sql_mode	Consistent with the read-write instance	No	This parameter controls SQL execution behavior in the analysis engine. Its behavior logic is consistent with that of MySQL. However, certain SQL_MODE values are either incompatible or configurable but not effective. For details, see the SQL_MODE section in Compatibility and Use Limits.
libra_enable_late_materialization	ON	No	Controls whether to enable or disable the lazy materialization feature. For more details, see Lazy Materialization.

tencent cloud

Prerequisites

Managing Parameters

Parameter List

Storage Engine Parameters

Row and Column Data Loading Parameters

Computing Engine Parameters