HD Insight 3.6: Data Factory V2- Running Script Actions to Install Dependencies for On-Demand Cluster

时间:2018-01-23 19:36:33

标签: azure apache-spark cluster-computing hdinsight azure-data-factory

I am currently trying to use script actions to deploy an On-Demand cluster (spark), using Data Factory V2, to install dependencies that my Python scripts need to run. However, the concept of script actions seems lost with V2.

How do I install packages on the cluster nodes as of now ? Is there an alternative ?

1 个答案:

答案 0 :(得分:0)

据我所知,似乎HD Insight仍支持脚本操作,用于安装外部依赖项:

  

从Azure门户,从启动板,单击Spark群集的磁贴(如果将其固定到起始板)。您还可以在浏览全部>下导航到您的群集。 HDInsight Clusters。

     

在Spark群集刀片中,单击左窗格中的“脚本操作”。运行在头节点和工作节点中安装TensorFlow的自定义操作。可以从以下引用bash脚本:https://hdiconfigactions.blob.core.windows.net/linuxtensorflow/tensorflowinstall.sh访问有关如何使用自定义脚本操作的文档。

完整文档可用here