Apache Kylin 1.5.4 发布,修复大量 Bug
欢迎加入运维网交流群:263444886Apache Kylin 1.5.4 正式发布了。
该版本包括漏洞修复,增强功能以及新的功能。它是 V1.5.3 改进版。升级之后,你仍然需要更新协处理器,点击查看如何更新协处理器
更新内容:
新特性
- Support Window Function
- UI for TopN: specify encoding and multiple “group by”
- Search cube by name in Web UI
- Collect Metrics to JMX
- Support Grouping Funtions
- Add a companion tool of CubeMetaExtractor for cube importing
Bug修复
- Cube Designer can’t drag rowkey normally
- Filter(CubeName) on Jobs/Monitor page works only once
- When modifying a model, Save after deleting a lookup table. The internal error will pop up.
- Save query hits org.apache.hadoop.hbase.TableNotFoundException: kylin_metadata_user
- unload non existing table cause NPE
- java.lang.IllegalArgumentException: Value not exists! - in Step 4 - Build Dimension Dictionary
- Consensus Problem when running the tool, MetadataCleanupJob
- Didn’t deal with the failure of renaming folder in hdfs when running the tool CubeMigrationCLI
- Error to load slow query in “Monitor” page for non-admin user
- Deploy in cluster mode, the “query” node report “scheduler has not been started” every second
- ‘Value not exist’ During Cube Merging Caused by Empty Dict
- Linkage error while executing any queries
- Models are missing after change project’s name
- Error handling for diagnosis
- Can’t query from child cube of a hybrid cube after its status changed from disabled to enabled
- Project name is always constant instead of real project name in email notification
- System Menu UI ACL issue
- Access denied when query seek to hybrid
- java.lang.NegativeArraySizeException when Build Dimension Dictionary
- CubeMigrationCLI: associate model with project
- CubeMigrationCLI: make global dictionary unique
- Clear ThreadLocal Contexts when query failed before scaning HBase
- Keep original column order when designing cube
- Job engine lock is not> - error start time at query result page
- Move all storage side behavior hints to GTScanRequest
改进
- Add Env and Project Info in job email notification
- The Key of the Snapshot to the>
- Should exclude those joins in whose>
- Remove all InvertedIndex(Streaming purpose)> - Add tip for field at ‘Add Streaming’ table page.
- Upgrade dependency libraries
- Make roaring bitmap version determined
- Upgrade to Avatica 1.8 or higher
- WebUI for GlobalDictionary
- Add more comments and default value for kylin.properties
- Support Separate HBase Cluster with NN HA and Kerberos Authentication
- Add view CubeInstance json function
- Improve the logic to decide whether to pre aggregate on Region server
- Add access controller to query
- Region server metrics: replace int type for long type for scanned row count
- Do not allow cross project clone for cube
- Loosen the constraint on FK-PK data type matching
- Improve enable limit logic (exactAggregation is too strict)
- Add owner for DataModel
- Show submitter for slow query
- BuildInFunctionTransformer should be executed per CubeSegmentScanner
- Delegate the loading of certain package (like slf4j) to tomcat’s parent> - Check duplicated measure name
- Refactor IJoinedFlatTableDesc
- Move hackNoGroupByAggregation to cube-based storage implementations
- Don’t use compression in packaging configuration
- SnapshotTable should only keep the columns described in tableDesc
- Add pivot feature back in query result page
- Make the creating intermediate hive table steps configurable (two options)
下载地址:
[*] https://www.apache.org/dyn/closer.cgi?path=/kylin/apache-kylin-1.5.4/
[*] https://kylin.apache.org/download/
发行说明:
[*] https://kylin.apache.org/docs15/release_notes.html
也可以查看邮件列表获取更多信息
页:
[1]