1.7.12 (Released on January 25th , 2021)
Overview
The version - V1.7.12 - of Indexima is now available.
This version offers new features :
Easy to use
Easy to adopt
Introducing in preview the Smart Tables. A smart table is a SQL table that has the capacity to perform analysis and index creation by itself! Our engine establishes the best moment to run, selects the most needed tables, estimates that the total indexation time won't exceed a threshold, and finally optimizes these tables.
Easy to implement
Useful links
- Download version 1.7.12
- Migrate from 1.7.11 to 1.7.12
- Procedure for upgrading Galactica (Indexima core)
- Procedure for upgrading Developer Console (VisualDoop2)
Release Note for the Indexima version 1.7.12
In Details
Core
- INDXMA-5400: Introducing the Smart Tables!
- INDXMA-3410: Index, tables, limited tables memory limit parameters can now be changed on-the-fly.
- INDXMA-5233: Analyzer can retrieve queries based on hour and minute, not only days, to give you better indexes.
- INDXMA-5282: Add a new parameter
spill.result.max_size.mb
to send an error when a spill query consumes too much disk space and slows the cluster. - INDXMA-5290: Add a new LOAD parameter named SUFFIX which sets a default value for columns after the last loaded column. This is useful when your data has fewer columns than the table in Indexima.
- INDXMA-5321: When a table is unloaded with the UNLOAD command, it won't be loaded at the next cluster reboot.
- INDXMA-5329: Add support for HASH and MD5 function.
- INDXMA-5368: Add support for beeline on Cloudera 6.3.
- INDXMA-5500: Increase performance by ~30% on LOAD / ADD INDEX command with COUNT(DISTINCT).
- INDXMA-5530: Increase drastically performance when doing an ADD INDEX from another index in memory, without using bucket nor prejoins. This optimization is enabled by default and can be disabled with the parameter
optimization.add_index.nb_rows
. - INDXMA-5563: Add a new parameter
load.feedback.rate
to define when Indexima will refresh load information - number of lines - in Monitor. - INDXMA-5655: Support Ascii function
External tables
- INDXMA-5284: Delegated queries now support the HAVING SQL keyword.
- INDXMA-5372: Add support for table creation on Sybase IQ.
- INDXMA-5446: Address an issue where we could not synchronize an external table, depending on JDBC columns types.
Governance
- INDXMA-4240: Add a new notification to display an external table without indexes.
- INDXMA-4718: You can now authenticate through LDAP group object.
- INDXMA-5280: Indexima now starts with
-XX:NewRatio=3
to have a better default heap memory management. - INDXMA-5320: Add an UNLOAD SCHEMA command to unload all tables in the specified schema, instead of doing it table by table.
- INDXMA-5339: Add a new notification to display index unused for 30 days.
- INDXMA-5354: Subqueries metrics are now stored in queries.csv to offer a better understanding of complex queries.
- INDXMA-5368: Add support for beeline on Cloudera 6.3.
- INDXMA-5387: Analyzer computes a better cardinality when using multiple expressions and calculated fields, leading to better index proposals.
- INDXMA-5548: The log4j2 template file has been updated for Hadoop 3 for better log filtering.
- INDXMA-5556: Add a new LDAP template for Developer Console to simplify the LDAP configuration.
Bug Fixes
- INDXMA-5298 (ZenD #867 #926): Address an issue where a TRUNCATE on an unloaded dimension table could lead to an error.
- INDXMA-5323: Address an issue where the migration from an older version could lead to incorrect aggregates values in indexes.
- INDXMA-5347 (ZenD #760): Address an issue where COUNT(DISTINCT) on a dimension table doesn't return any value instead of 0.
- INDXMA-5385 (ZenD #893 #921): Address an issue where a table could be unloaded on creation when the cluster has low memory.
- INDXMA-5475 (ZenD #966): Address an issue where SUM(IF) optimization could lead to SELECT queries on error.
- INDXMA-5486: Address an issue where query history could be duplicated.
- INDXMA-5490 (ZenD #971): UNION queries could fail when using an indexed CASE WHEN clause.
- INDXMA-5504: Global memory limit was still incremented during a query kill, leading to refusing more queries than necessary.
- INDXMA-5512 (ZenD #994): Address an issue where the split bucket happens too much, leading to an error.
- INDXMA-5557 : Address an issue where multiple joins, especially CROSS JOIN, could lead to an error.
- INDXMA-5602 / 5603 (ZenD #1035 #1067) : Address some issues when doing multiple UNION queries.