Miscellaneous Parameters
enable_cluster_resize
Parameter description: If an SQL statement involves tables belonging to different groups, you can enable this parameter to push the execution plan of the statement to improve performance.
Type: SUSET
Value range: Boolean
- on indicates the execution plan of the statement can be pushed.
- off indicates the execution plan of the statement cannot be pushed.
Default value: off
This parameter is used for internal O&M. Do not set it to on unless absolutely necessary.
cstore_insert_mode
Parameter description: Specifies the storage location of data to be imported to an HDFS table. This parameter is needed for operations that involve data import, such as INSERT, UPDATE, COPY, and VACUUM FULL.
Type: USERSET
Value range: enumerated values
- AUTO: The major part of imported data is stored in HDFS, and the tail is stored in the delta table.
- DELTA: All the imported data is stored in the delta table.
- MAIN: All the imported data is stored in HDFS.
Default value: auto
You can set other values as the default in the configuration file.
remote_read_mode
Parameter description: When enable_crc_check is set to on and the data read by the primary DN fails the verification, remote_read_mode is used to specify whether to enable remote read and whether to use secure authentication for connection upon the data verification failure. The setting takes effect only after the cluster is restarted.
Type: POSTMASTER
Value range: off, non_authentication, authentication
- off: indicates that the remote read function is disabled.
- non_authentication: indicates that the standby DN is connected and data is obtained when non-authentication is used.
- authentication: indicates that the standby DN is connected and data is obtained through authentication. Before restarting the cluster, ensure that a certificate exists in the $GAUSSHOME/share/sslcert/grpc/ directory. Otherwise, the cluster cannot be started.
Default value: non_authentication
enable_upgrade_merge_lock_mode
Parameter description: If this parameter is set to on, the delta merge operation internally increases the lock level, and errors can be avoided when update and delete operations are performed at the same time.
Type: USERSET
Value range: Boolean
- If this parameter is set to on, the delta merge operation internally increases the lock level. In this way, when any two of the DELTAMERGE, UPDATE, and DELETE operations are concurrently performed, an operation can be performed only after the previous one is complete.
- If this parameter is set to off, and any two of the DELTAMERGE, UPDATE, and DELETE operations are concurrently performed to data in a row in the delta table of the HDFS table, errors will be reported during the later operation, and the operation will stop.
Default value: off
job_queue_processes
Parameter description: Specifies the number of jobs that can be concurrently executed. This parameter is a postmaster parameter. You can set it using gs_guc, and you need to restart gaussdb to make the setting take effect.
Type: POSTMASTER
Value range: 0 to 1000
Functions:
- Setting job_queue_processes to 0 indicates that the scheduled task function is disabled and that no job will be executed. (Enabling scheduled tasks may affect the system performance. At sites where this function is not required, you are advised to disable it.)
- Setting job_queue_processes to a value that is greater than 0 indicates that the scheduled task function is enabled and this value is the maximum number of tasks that can be concurrently processed.
After the scheduled task function is enabled, the job_scheduler thread at a scheduled interval polls the pg_jobs system catalog. The scheduled task check is performed every second by default.
Too many concurrent tasks consume many system resources, so you need to set the number of concurrent tasks to be processed. If the current number of concurrent tasks reaches job_queue_processes and some of them expire, these tasks will be postponed to the next polling period. Therefore, you are advised to set the polling interval (the interval parameter of the submit interface) based on the execution duration of each task to avoid the problem that tasks in the next polling period cannot be properly processed because overlong task execution time.
Note: If the number of parallel jobs is large and the value is too small, these jobs will wait in queues. However, a large parameter value leads to large resource consumption. You are advised to set this parameter to 100 and change it based on the system resource condition.
Default value: 10
ngram_gram_size
Parameter description: Specifies the length of the ngram parser segmentation.
Type: USERSET
Value range: an integer ranging from 1 to 4
Default value: 2
ngram_grapsymbol_ignore
Parameter description: Specifies whether the ngram parser ignores graphical characters.
Type: USERSET
Value range: Boolean
- on: Ignores graphical characters.
- off: Does not ignore graphical characters.
Default value: off
ngram_punctuation_ignore
Parameter description: Specifies whether the ngram parser ignores punctuations.
Type: USERSET
Value range: Boolean
- on: Ignores punctuations.
- off: Does not ignore punctuations.
Default value: on
zhparser_dict_in_memory
Parameter description: Specifies whether Zhparser adds a dictionary to memory.
Type: POSTMASTER
Value range: Boolean
- on: Adds the dictionary to memory.
- off: Does not add the dictionary to memory.
Default value: on
zhparser_multi_duality
Parameter description: Specifies whether Zhparser aggregates segments in long words with duality.
Type: USERSET
Value range: Boolean
- on: Aggregates segments in long words with duality.
- off: Does not aggregate segments in long words with duality.
Default value: off
zhparser_multi_short
Parameter description: Specifies whether Zhparser executes long words composite divide.
Type: USERSET
Value range: Boolean
- on: Performs compound segmentation for long words.
- off: Does not perform compound segmentation for long words.
Default value: on
zhparser_multi_zall
Parameter description: Specifies whether Zhparser displays all single words individually.
Type: USERSET
Value range: Boolean
- on: Displays all single words separately.
- off: Does not display all single words separately.
Default value: off
zhparser_multi_zmain
Parameter description: Specifies whether Zhparser displays important single words separately.
Type: USERSET
Value range: Boolean
- on: Displays important single words separately.
- off: Does not display important single words separately.
Default value: off
zhparser_punctuation_ignore
Parameter description: Specifies whether the Zhparser segmentation result ignores special characters including punctuations (\r and \n will not be ignored).
Type: USERSET
Value range: Boolean
- on: Ignores all the special characters including punctuations.
- off: Does not ignore all the special characters including punctuations.
Default value: on
zhparser_seg_with_duality
Parameter description: Specifies whether Zhparser aggregates segments in long words with duality.
Type: USERSET
Value range: Boolean
- on: Aggregates segments in long words with duality.
- off: Does not aggregate segments in long words with duality.
Default value: off
acceleration_with_compute_pool
Parameter description: Specifies whether to use the computing resource pool for acceleration when OBS is queried.
Type: USERSET
Value range: Boolean
- on indicates that the query covering OBS is accelerated based on the cost when the computing resource pool is available.
- off indicates that no query is accelerated using the computing resource pool.
Default value: off
behavior_compat_options
Parameter description: Specifies database compatibility behavior. Multiple items are separated by commas (,).
Type: USERSET
Value range: a string
Default value: In upgrade scenarios, the default value of this parameter is the same as that in the cluster before the upgrade. When a new cluster is installed, the default value of this parameter is check_function_conflicts to prevent serious problems caused by incorrect function attributes defined by users.
- Currently, only Table 1 is supported.
- Multiple items are separated by commas (,), for example, set behavior_compat_options='end_month_calculate,display_leading_zero';
- strict_concat_functions and strict_text_concat_td are mutually exclusive.
table_skewness_warning_threshold
Parameter description: Specifies the threshold for triggering a table skew alarm.
Type: SUSET
Value range: a floating point number ranging from 0 to 1
Default value: 1
table_skewness_warning_rows
Parameter description: Specifies the minimum number of rows for triggering a table skew alarm.
Type: SUSET
Value range: an integer ranging from 0 to INT_MAX
Default value: 100000
max_cache_partition_num
Parameter description: Specifies the number of memory-saving partitions in column-store mode during redistribution after scale-out. If the number of partitions exceeds the upper limit, the earliest cached partition is directly written to the column-store file.
Type: SIGHUP
Value range: an integer ranging from 0 to 32767.
- 0 indicates that the memory-saving mode is disabled in column storage.
- Values from 1 to 32767 indicate the maximum number of partitions that can be cached in a partitioned table.
Default value: 0
This parameter is used for redistribution during scale-out. A proper value can reduce the memory consumption during redistribution of a partitioned column-store table. However, tables with unbalanced data distribution in some partitions may generate a large number of small CUs after the redistribution. If there are a large number of small CUs, execute the VACUUM FULL statement to merge them.
enable_prevent_job_task_startup
Parameter description: Specifies whether to prevent the thread startup of scheduled jobs. This is an internal parameter. You are not advised to change the value of this parameter.
Type: SIGHUP
Value range: Boolean
- on: Threads of scheduled jobs will not be started.
- off: Threads of scheduled jobs will be started.
Default value: off
Set this parameter only on CNs.
auto_process_residualfile
Parameter description: Specifies whether to enable the residual file recording function.
Type: SIGHUP
Value range: Boolean
- on indicates that the residual file recording function is enabled.
- off indicates that the residual file recording function is disabled.
Default value: off
enable_view_update
Parameter description: Enables the view update function or not.
Type: POSTMASTER
Value range: Boolean
- on indicates that the view update function is enabled.
- off indicates that the view update function is disabled.
Default value: off
view_independent
Parameter description: Decouples views from tables, functions, and synonyms or not. After the base table is restored, automatic association and re-creation are supported.
Type: SIGHUP
Value range: Boolean
- on indicates that the view decoupling function is enabled. Tables, functions, synonyms, and other views on which views depend can be deleted separately (except temporary tables and temporary views). Associated views are reserved but unavailable.
- off indicates that the view decoupling function is disabled. Tables, functions, synonyms, and other views on which views depend cannot be deleted separately. You can only delete them in the cascade mode.
Default value: off
bulkload_report_threshold
Parameter description: Sets the threshold for reporting import and export statistics.
Type: SIGHUP
Value range: an integer ranging from 0 to INT_MAX
Default value: 50
assign_abort_xid
Parameter description: Determines the transaction to be aborted based on the specified XID in a query.
Type: USERSET
Value range: a character string with the specified XID
This parameter is used only for quick restoration if a user deletes data by mistake (DELETE operation). Do not use this parameter in other scenarios. Otherwise, visible transaction errors may occur.
Feedback
Was this page helpful?
Provide feedbackThank you very much for your feedback. We will continue working to improve the documentation.See the reply and handling status in My Cloud VOC.
For any further questions, feel free to contact us through the chatbot.
Chatbot