[Release] OrientDB 2.2.25 is out!

47 views
Skip to first unread message

Luca Garulli

unread,
Aug 2, 2017, 12:20:02 PM8/2/17
to OrientDB

2.2.25 - August 02, 2017

Functionality Added or Changed

Studio

  • A new 'Neo4j to OrientDB Importer' plugin that connects to Neo4j via the Neo4j remote bolt protocol has now been integrated into both OrientDB Community and Enterprise Editions. This is in addition to the already existing command-line 'Neo4j to OrientDB Importer" tool, that made use of embedded connections via the Java API (and that is still available as an external package) - Issue #432

Distributed

  • Improved java client reconnection when Data Centers are used: two new values for the -Dnetwork.retry.strategy option have now been added: same-dc and autoIssue #7509Documentation

ETL

  • Added support for algorithm parameter in index creation - Issue #7581

Bugs Fixed

Core

  • Fixed Bug on JSON import in multithreading environment - Internal Issue #9020
  • Fixed bug with database backup/restoration and cache state file, cache state file is now ignored - Issue #7601 & Internal Issue #9024
  • Improved performance of rid bag compare - Internal Issue #8803

SQL

  • Fixed query parsing issue when "in" clause occurs after "or" condition with parentheses - Issue #6834
  • Fixed exception in select from collection of DATETIME or DATE, using prepared statement - Issue #7574
  • Fixed comparison of Short values with IN and CONTAINS in SQL - Issue #7578
  • Fixed StackOverflow exception when a list of enums is passed as an SQL param - Issue #7418
  • Fixed usage of include function in Match statement - Issue #271
  • Fixed security checks in SQL to allow reading of own OUser record - Issue #7549

Server

  • Implemented GZip Compression for static resources - Issue #476

Distributed

  • When executing the HA SYNC CLUSTER command, part of the cluster was not migrated. Now fixed - Internal Issue #8982
    • Warning: HA SYNC CLUSTER <cluster-name> will now trigger an automatic rebuild of the indices defined on the Class the cluster <cluster-name> belongs to. Depending on the number of defined indices and amount of data included in the Class this operation can take some time. Furthermore it is important to wait that the sync cluster operation has finished before performing any database operations that involve that cluster.
  • Fixed issue related to local DC quorum management - Issue #7539 & Internal Issue #8981
  • Enlarged default timeout for database chunk deployment - Internal Issue #8923
  • Fixed issue that was causing record version to be different on different nodes - Internal Issue #9003
  • Fixed issue with distributed sequences

Lucene Spatial

  • Fixed lifecycle on the legacy spatial index - Issue #7585

Console

  • Fixed retrieval of custom conflict strategy from remote - Internal Issue #8931



Best Regards,

Luca Garulli
Founder & CEO
Reply all
Reply to author
Forward
0 new messages