在Datapipeline使用自定义的AMI（带s3cmd） [英] Using a custom AMI (with s3cmd) in a Datapipeline

问题描述

如何安装用于在管道上的AMI s3cmd？ 这应该是一个相当基本的事情，但我似乎不能把它做： 下面是我尝试过：

How can I install s3cmd on a AMI that is used in the pipeline? This should be a fairly basic thing to do but I can't seem to get it done: Here's what I've tried:

  
• 在启动时没有图像-id选项管道=>一切工作正常
  
• 导航到EC2和创建的运行实例的图像，以确保所有需要的东西在管道中运行安装在我的自定义AMI
  
• 手工启动这个AMI作为一个实例
  
• 通过SSH连接到机器并安装S3cmd
  
• 创建本机的另一图像，安装这次s3cmd
  
• 关闭实例
  
• 又开始酝酿中，这一次与新创建的AMI作为图像ID和S3cmd安装
  

• Started a Pipeline without the Image-id option => Everything works fine
• Navigated to EC2 and created an Image of the running Instance to make sure all the needed stuff to run in the pipeline is installed on my custom AMI
• Started this AMI manually as an Instance
• SSH'd into the machine and installed S3cmd
• Created another Image of the machine, this time with s3cmd installed
• Shut down the Instance
• Started the Pipeline again, this time with the newly created AMI as Image-id and S3cmd installed

现在的资源开始跑，但我的活动（ShellCommandActivity）卡在WAITING_FOR_RUNNER状态和脚本从来没有被执行。

Now the Resource starts "RUNNING" but my Activity (ShellCommandActivity) is stuck in the WAITING_FOR_RUNNER state and the script never gets executed.

我有什么做的就是使用自定义映像运行的管道？还是甚至有一个更简单的方法来使用s3cmd在管道？

What do I have to do to get the pipeline running with a custom image? Or is there even an easier way to use s3cmd in a pipeline?

感谢您！

推荐答案

我想通了，现在，通过使用一个干净亚马逊的Linux AMI（例如从市场上），并安装S3cmd，而不是创建一个AMI出正在运行的管道资源。我看到这个AMI有不同的内核版本，所以这可能是这个问题。

I figured it out now, by using a "clean" Amazon Linux AMI (from the marketplace for example) and installing S3cmd, rather than creating an AMI out of a running Pipeline Resource. I saw that this AMI has a different Kernel version, so this could have been the problem.

这篇关于在Datapipeline使用自定义的AMI（带s3cmd）的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！