BI/Spider Parameter

6.0

/

6.0.2

Optimized the logic of identifying data types to judge the number of data bits accurately.

6.0.3

Added Intelligent Data Explanation under BI Parameter.

Added Extract query timeout (seconds) under BI Parameter.

Renamed Spider Parameter as Extracted Parameter.

6.0.4

1. Added new parameters.

Added Excel export remaining disk space(G) under BI Parameter.

Added Update High Performance Time Interval under Extracted Parameter.

2. Deleted parameters.

Deleted Spark Driver Port under Extracted Parameter.

Deleted Spark blockManager Port under Extracted Parameter.

Deleted Spark Adaptive Adjustment under Extracted Parameter.

3. Optimized parameter classification and adjusted the classification of some parameters.

Moved Limit on number of threads during detail table export and Threads for multiple indicator calculation to the category of Advanced Tuning Parameters under BI Parameter.

Moved Spark log output level and Date simple load mode to the category of Basic Parameters under Extracted Parameter.

Moved Spark local temporary file directory and Incremental update data block plan to the category of Advanced Tuning Parameters under Extracted Parameter.

6.0.5

Intelligent Data Explanation under BI Parameter is disabled by default.

6.0.8

1. Renamed Extracted Parameter as Spider Parameter.

2. Added new parameters.

Added Data access active period(days) under Spider Parameter.

Added Data creation active period (days) under Spider Parameter.

Added Data activity global check time under Spider Parameter.

3. Deleted parameters.

Deleted Analyze user self-service dataset disk occupation size (cell) under Spider parameter.

6.0.9

Added restrictions on unreasonable calculations. If the data volume is too huge after Group Summary or Left and right merge, the calculation in components/dashboards will be interrupted.

Data Type Identification

1. Regardless of whether it is enabled or not, a field with decimal places is always identified as a value type.

2. When it is disabled, a field with over 15 digits is identified as a text type, while that with 15 digits or less is value.

3. When it is enabled, the fields are identified as value types (double), with valid precision of 15 digits (2^52). Therefore, the fields with over 15 digits may cause precision loss.

4. Direct and distributed tables added without enabling the parameter:

• If a table has not been edited, after enabling the parameter and restarting FineBI, you will find the fields display as value types when going to the editing interface. Just save it.

• If a table has been edited (Added the field type conversion function in FineBI5.1.5 and later versions), after enabling the parameter and restart FineBI, you will find the fields still display as text types instead of value types when going to the editing interface.

Disable

Restart

SQL parameter widget filtering takes effect

Whether the widget binding SQL parameter function and filtering function take effect at the same time.

Disabled by default, indicating no effect.

Disable

Not

Chinese Order

Whether to use Chinese order.

Disabled by default, indicating not using Chinese order.

After enabling it, extract tables need to be re-extracted.

Disable

Restart

Amount of Data Access

Amount of data access limits the number of data rows that can be read into server memory. Not all calculation scenarios with large   data volume read all data into memory. FineBI’s Spider engine has intelligent memory usage strategies, such as:

• If the number of data rows after grouping and summarizing exceeds the set value, the calculation in the     component/dashboard will be interrupted.

• For smaller data tables when performing Left and right merge, if their associated fields after grouping exceeds the set value, the calculation in the component/dashboard will be interrupted.

Low configuration may affect the accuracy of data calculation.

High configuration may lead to system crash.

Unit: row

Range: 0~10,000,000

Recommended range: 10,000~1,000,000. Recommend   keeping the default value.

1000000

Restart

Data volume limit on Excel export

If the data volume limit is exceeded when you export   dashboards to Excel, an error message will occur.

Unit: cell (No. of row * column)

Range: 0~2,000,000,000

Recommended range: 0~1,000,000,000

Null

Not

Excel export remaining disk space(G)

When you export Excel, a disk space check is performed. If the disk space is less than the limit value, the export will be   interrupted, which can avoid system crash caused by full disk.

If the limit value is exceeded, an error message will occur: The current exported data volume is too large and there is not enough disk space. The system protection mechanism has been triggered, and the export is interrupted. Please clear the disk space in time, limited by [Excel Export Disk Space Check] parameter, disk space limit: {number} G.

Range: 5~unlimited

5

Not

Direct query timeout(seconds)

Long component query time caused by too many   components, and blocked FineBI requests caused by long query time for a component may be mistaken as crash.

To solve the problem, you can set the direct query timeout value. Any real-time data query will be interrupted when it times out in case that slow queries block normal queries.

An error massage will occur: Component query time exceeds {number} minutes, query is interrupted.

Unit: seconds

Recommended range: 10~300

Effect scope: FineBI direct query requests for anything except table structure

180

Not

Intelligent Data Explanation

If you do not configure components, Intelligent Data   Explanation is disabled by default.

This configuration takes effect only after installing the Data Explanation plugin.

Disabling it can trigger manually configured Data Explanation to avoid abuse of data interpretation.

Disable

Not

Extract query timeout (seconds)

Long component query time caused by too many   components, and blocked FineBI requests caused by long query time for a component may be mistaken as crash.

To solve the problem, you can set the extract query timeout value. Any extract data query will be interrupted when it times out in case that slow queries block normal queries.

An error message will occur: Component query time exceeds {number} minutes, query is interrupted.

Unit: seconds

Recommended range: 10~300

45

Not

Limit on number of threads during detail table   export

When multiple users export Excel with large data volume at the same time, the number of users may exceed that of thread, thus affecting   user experience. Therefore, you can use the parameter to set the number of the exporting users. If the limit value of users exporting detail tables is exceeded, you need to wait.

Range: 1~10

Recommended range: 1~5. Recommend keeping the default value.

3

Restart

Default Update Settings for Self-service Dataset

Whether self-service dataset single table updates   with the parent table.

For details, see Update of Self-service Dataset   Single Table.

Update with the Parent Table

Restart

Spark log output level

Spark log output level is used to define log level.   Log uses standard output stream, and outputs into the catalina.out file of Tomcat or the nohup file of FineBI.

Options: INFO, WARN, ERROR, DEBUG.

• INFO: print error messages and basic execution logs.

• WARN: print warning or prompt messages.

• ERROR: only print error logs.

• DEBUG: print all logs.

INFO

Restart

Update High Performance Time Interval

There are no restrictions on the number of update   resource pool cores occupied during the configuring time period.

In order not to affect users, it is recommended to   configure a time period when the system is unused and update tasks are frequent.

Null

Not

Date simple load mode

When the simple load mode is enabled, the date field only generates a few group types in advance when extracted, which speeds up the generation and reduces occupied space.

Ungenerated groups may encounter performance loss when computed.

Disable

Restart (Need to update data after restart)

Data access active period(days)

The continuous active period after the dataset is   accessed.

If admins set Data access active period(days) as n days, the dataset under two scenarios will be seen as in an active state:

• Dashboards or components created through the dataset have been edited or viewed within n days.

• Its child table is edited and saved within n days.

Unit: days

14

Restart

Data creation active period (days)

The continuous active period after the dataset is   created.

A dataset is considered in an active state within n   days of its creation, and inactive beyond that.

Unit: days

7

Restart

Data activity global check time

The check time for judging the active state of the   dataset, at which time the access and creation of the dataset will be checked uniformly, and its active state will be refreshed, so that some datasets will not be extracted.

You can select from the drop-down list:

Fixed time check: check every day at {hour} {minute} {second}, with the format of hh:mm:ss

Check before global update scheduled tasks

Fixed time check

00:00:00

Restart

Compressed slice thread count

Number of slice threads (compression & writing)   when data is extracted

If the memory is limited (4 GB or less) and cannot be expanded, you can reduce the thread count to alleviate memory pressure.

Low configuration will cause slower speed of data update.

For high configuration, other tasks may be influenced when data is updated.

Recommend keeping the default value of 8. Recommended Interval: 4~16

8

Restart

Queue length of unprocessed waiting slices

The queue length of unprocessed waiting slices when data is extracted.

If the memory is limited (4 GB or less) and cannot   be expanded, you can reduce the queue length to alleviate memory pressure.

Low configuration will cause slower speed of data   update.

For high configuration, other tasks may be influenced when data is updated.

Recommend keeping default value of 200.

Recommended range: [100~200]

200

Restart

The number of threads executed for the new data   extraction task

The number of threads that the extraction tasks for new data are simultaneously executed

Low configuration will cause slower speed of data   update.

For high configuration, other tasks may be influenced when data is updated.

Recommend keeping default value of 5.

Recommended Interval: [1~5]

5

Restart

Spark local temporary file directory

Directory for Spark to write temporary files

It needs some space. Setting to the SSD mount path can improve the performance of Spark processing association and Spark SQL query.

Note: This parameter is invalid in cluster version of FineBI, and needs to be configured on the server end.

Null (tmp under Linux)

Restart

Incremental update data block plan

During the set time period, incremental update tasks will not be merged, which improves the speed of incremental update.

Format: hh:mm:ss – hh:mm:ss

Example: 10:10:10 – 12:12:12

Disable

Restart