共用方式為


適用於 Apache Spark 3.4 的 Azure Synapse Runtime (EOSA)

警告

終止支援公告

  • 針對適用於 Apache Spark 3.4 的 Azure Synapse Runtime 宣佈終止支援。
  • 適用於 Apache Spark 3.4 的 Azure Synapse Runtime 將會淘汰並停用 2026 年 3 月 31 日。 

Azure Synapse Analytics 支援 Apache Spark 的多個運行時間。 本文件涵蓋適用於 Apache Spark 3.4 的 Azure Synapse Runtime 的運行時間元件和版本。

元件版本

元件 版本
Apache Spark 3.4.1
作業系統 水手 2.0
JAVA 11
程式語言 Scala 2.12.17
三角洲湖 2.4.0
Python 3.10
R 4.2.2

提示

如需最新資訊、變更的詳細清單,以及Spark運行時間的特定版本資訊,請檢查並訂閱 Spark運行時間版本和更新

程式庫

若要檢查適用於 Java/Scala 的 Azure Synapse Runtime 中所包含的連結庫,Python 和 R 會移至 適用於 Apache Spark 3.4 的 Azure Synapse Runtime 版本資訊

提示

spark.memoryOverheadFactor.preferred:若設為 true,Spark 會優先將 spark.driver.memoryOverheadFactor 與 spark.executor.memoryOverhead 置於 spark.driver.memoryOverhead 與 spark.executor.memoryOverhead 明確定義的值之上。 啟用後,開銷總是透過因子計算,且會忽略明確的開銷設定。 預設為 false,以支援向下相容。