大的架构更改：在开发服务器上执行此操作有意义吗？
基本上：如果我使用批量上传器上传从我的开发服务器生成的一堆数据，关键引用会保持一致吗？
精心设计：
我想做一个大的架构更改。每天晚上我有六个小时的时间几乎为零活动，在此期间我可以将应用程序设置为只读模式。
我真的不想弄清楚如何进行纠错、映射缩减等。由于我可以将应用程序关闭一段时间，因此使用批量上传器下载数据并在我的开发人员上处理它是否有意义服务器，并上传新数据？
我会很高兴最终找到 MapReduce 来清理周围的死数据。我的所有实体种类名称都在发生变化，因此暂时保留一堆旧的、过时的数据不会成为问题。
Basically: if I upload a bunch of data generated from my dev server with the bulk uploader, will key references stay consistent?
Elaboratedly:
I have a big schema change I want to do. Every night I have a six-hour period with almost zero activity, during which I could set the app to read-only mode.
I don't really want to deal with figuring out how to do error correction, mapreduce, etc. Since I can take the app down for a while, does it make sense to download the data with the bulk uploader, process it on my dev server, and upload the new data?
I'll be happy enough to figure out MapReduce eventually to clean up the dead data lying around. All of my entity kind names are changing, so having a bunch of old, outdated data around for a while won't be a problem.
如果你对这篇内容有疑问，欢迎到本站社区发帖提问 参与讨论，获取更多帮助，或者扫码二维码加入 Web 技术交流群。
绑定邮箱获取回复消息
由于您还没有绑定你的真实邮箱，如果其他用户或者作者回复了您的评论，将不能在第一时间通知您！
发布评论
评论（1）
当然，如果您愿意，您当然可以这样做。在不知道更改的性质的情况下，很难说这是否比编写 MapReduce 来更容易。
一个更简单的选择可能是使用remote_api来修改数据，而不需要显式转储和恢复它。与本地计算机之间的传输相同，但不需要额外的步骤。
You can certainly do this if you want, of course. Without knowing the nature of the changes, it's hard to say if this will be any easier than writing a mapreduce to do it.
An easier option would probably be to use remote_api to modify the data, without explicitly dumping and restoring it. The same transfer to and from your local machine happens, but without the extra steps needed.