2019 年 8 月August 2019

这些功能和 Azure Databricks 平台的改进功能已于 2019 年 8 月发布。These features and Azure Databricks platform improvements were released in August 2019.

备注

发布分阶段进行。Releases are staged. 在初始发布日期后,可能最长需要等待一周,你的 Azure Databricks 帐户才会更新。Your Azure Databricks account may not be updated until up to a week after the initial release date.

新区域:阿拉伯联合酋长国北部New region: UAE North

2019 年 8 月 27 日Aug 27, 2019

Azure Databricks 现已在阿拉伯联合酋长国北部(迪拜)推出。Azure Databricks is now available in UAE North (Dubai).

工作区库安装增强Workspace library installation enhancement

2019 年 8 月 27 日至 9 月 3 日:版本 3.1Aug 27 - Sept 3, 2019: Version 3.1

管理员现在可以通过群集的“库”选项卡在所有群集上安装库。Admins can now install a library on all clusters from a cluster’s Libraries tab.

安装群集库Install cluster library

有关详细信息,请参阅工作区库For details, see Workspace library.

群集 UI 现在会反映更一致的交互式和自动化群集术语Clusters UI now reflects more consistent interactive and automated cluster terminology

2019 年 8 月 27 日至 9 月 3 日:版本 3.1Aug 27 - Sept 3, 2019: Version 3.1

为了明确你为作业使用哪一种类型的群集(自动或交互式),“群集”列表和作业配置群集 UI 进行了更新。To clarify which type of cluster (automated or interactive) you are using for a job, the Clusters list and job Configure Cluster UI have been updated.

在“群集”列表中,“作业群集”子列表已重命名为“自动群集”,以反映为作业创建的群集按自动费率收费这一事实。In the Clusters list, the Job Clusters sublist has been renamed to Automated Clusters to reflect that a cluster created for a job is charged at the automated rate.

自动群集Automated clusters

为作业配置群集时,“群集类型”菜单选项“新群集”已重命名为“新自动群集”,“现有群集”已重命名为“现有交互式群集”。When you configure a cluster for a job, the Cluster Type menu option New Cluster has been renamed to New Automated Cluster and Existing Cluster has been renamed to Existing Interactive Cluster. 这些更改反映了这样一个事实:如果你选择新群集,它将按自动费率收费;如果你选择现有群集,它将按交互费率收费。These changes reflect the fact that if you choose a new cluster it is charged at the automated rate, and if you choose an existing cluster it is charged at the interactive rate.

作业自动群集Job automated clusters

Databricks Runtime 5.5 和 Databricks Runtime 5.5 ML 现已推出 LTS 版本Databricks Runtime 5.5 and Databricks Runtime 5.5 ML are LTS

2019 年 8 月 27 日Aug 27, 2019

Databricks Runtime 5.5 和 Databricks Runtime 5.5 ML 被确定为长期支持 (LTS) 运行时。Databricks Runtime 5.5 and Databricks Runtime 5.5 ML have been identified as long term support (LTS) runtimes. Azure Databricks 为 LTS 运行时提供整整两年的支持。Azure Databricks provide two full years of support for LTS runtimes. 有关详细信息,请参阅 Databricks Runtime 发行说明For details, see Databricks runtime release notes.

针对池的实例分配通知Instance allocation notifications for pools

2019 年 8 月 13 日至 20 日:版本 2.104Aug 13 - 20, 2019: Version 2.104

现在,如果池未能分配最小空闲实例数,则会通知你。You are now informed if a pool fails to allocate the minimum number of idle instances. 池详细信息和统计信息页会显示一条错误消息,指出失败的实例数。The pool details and statistics page displays an error message stating the number of failed instances. 你可以单击错误消息以查看有关每个挂起实例失败的详细信息。You can click the error message to view details about each pending instance failure.

新的群集事件New cluster events

2019 年 8 月 13 日至 20 日:版本 2.104Aug 13 - 20, 2019: Version 2.104

群集事件现在会在群集固定或取消固定时生成。Cluster events are now generated whenever a cluster is pinned or unpinned.

请参阅 ClusterEventTypeSee ClusterEventType.

MLflow 更新MLflow updates

2019 年 8 月 13 日至 20 日:版本 2.104Aug 13 - 20, 2019: Version 2.104

笔记本中的“运行”边栏现在支持按指标或参数对运行排序。The Runs sidebar in notebooks now supports sorting the runs by metric or parameter.

MLflow 1.2 已于 2019 年 8 月 12 日发布。MLflow 1.2 was released on Aug 12, 2019. 有关详细信息,请参阅 MLflow 1.2 ChangelogFor details, see the MLflow 1.2 Changelog.