Releases: orientechnologies/orientdb
Releases · orientechnologies/orientdb
3.2.15
Core
- give more clear error when view metadata are not correct on view creation
- improved validation of embedded types to avoid use of vertices and edges in embedded
- handle additional case for shape type conversion in query engine
- removed not needed synchronization in views structures
- reduced logging of the security module
- minor fix for no-tx changes on lucene indexes
- removed duplicate logic in index implementation
- reduced the thread local lookups and database isClosed checks
- removed some locale translate done only on specific use cases
- avoid to update the index manager when there are no changes
- replaced storage error state lock with atomic reference
- removed server and client dependency from lucene module
Distributed
- minor fix on view creation and refresh in case of distributed environment
Release 3.2.14
Core
- Optimization on graph delete operations (issue #9821)
- Fix on remote deserialization of specific exception
- Fix on index logic for multi-page split
- Fix limit of index key size, now it fail with a user message when the key is over the maximum size (10Kb)
- Optimization in index write
- Optimization of security checks in writes
- Fix wrongly oversize WAL records
- Fix views locking logic to update the view atomically
- Fix failing schedule of view if previous view fail to refresh
- Add support of indexing of embedded types in views
- Fixed script poll release issue
Release 3.2.13
- Release Notes: https://github.com/orientechnologies/orientdb/wiki/OrientDB-3.2-Release-Notes#3213---1-December-2022
Core
- Make sure to use high level transactions in all index change cases
- Made index rebuild batch changes in transaction
- Fix view unlocking in case of refresh error
- Optimized update of vertices that use ridbag trees
- Fix regression on OResult hash code, issue #9900
- Ignore index corrupted errors on drop
- Batched view refreshing in transaction
- Optimized of usage of read/write lock in storage
- Fix check of index type also on chained indexes
- Reduced usage of modifiable schema to reduce contention on read/write lock
- Fix waiting for last task in ViewManager close
- Fix running of ddl while incremental backup is running do not block anymore other write operations.
- Fix invalid cluster id do not make storage readonly
Distributed
- Fix in handling of node missing on refresh
3.2.12
Released 3.2.12 hotfix with improvements in query performances, stabilization of the profiling in the enterprise edition and a good set of other fix in the console, core, query engine, and distribited.
Core
- Add timeout check in next query page fetching
- Updated dependencies
- Optimized Security Policies checks
- Optimized query engine property fetching and filtering
- Removed not used anymore dependencies
- Fix in truncate clusters logic
- Improved string caching logic
- Made the direct memory pointer soft reference to retain it and reduce the allocation overhead
- Add a correct error message when opening a not existing database in a in memory context
- Fix regression in sql function execution after refactor, issue #9888
- Optimization in storage record reading
Distributed
- Fix OClass.setCustom() in distributed setup
Console
- Fix regression database creation, issue #9889
Enterprise
- Limited the maximum number of metrics kept in memory
- Made query statements metrics ognostic of values