允许的最大内存分配所有层的例外情况 [英] Maximum allowable memory allocation Exceptions at all tiers

问题描述

我正在尝试将21GB数据库（大约200M记录）加载（处理）到Azure中的Tabular AS模型。鉴于公布的数据压缩值，我曾希望它能够"适应"压缩值。进入"S0"定价等级或"S1"等等。最糟糕的情况。除了
之外，我的理解是我不能使用"B"。因为我的源数据是> 16GB。



S0提供10GB缓存，而S2缓存为25GB。但是在这两种情况下我都得到了"内存错误：您已达到层的最大允许内存分配。考虑升级到具有更多可用内存的层"通过处理的方式大约三分之二
。即使它超出我的预算，我尝试升级到S3层（50 GB），但仍然遇到相同的错误。我已经阅读了muitiple网站/帖子但是找不到任何关于Azure上的SSAS如何使用内存的确定性。关于它是如何工作和/或我做错的任何
指导。

解决方案

Azure SSAS是一个内存数据库，因此您需要足够的内存来存储要分析的数据。听起来您拥有的数据超过了您购买的
等级目前所拥有的数据。正如消息所说，您可以升级到更昂贵的层，这将允许您处理更多数据;您还应该查看正在加载的数据并删除任何列，因为这会减少您的内存使用量。

您可以从Azure门户或Azure服务器属性中查看内存使用情况。

Analysis Services提供的指标可帮助您监控服务器的性能和运行状况。例如，监视内存和CPU使用情况，客户端连接数，
和查询资源消耗。 Analysis Services使用与大多数其他Azure服务相同的监视框架。要了解详情，请参阅  Metrics
in Microsoft Azure 和
分析服务的新内存选项。

有关详细信息，您可以参考

SO 解决类似问题的线程。

希望这会有所帮助

I am trying to load (Process) a 21GB data base (approx 200M records) to a Tabular AS Model in Azure. Given published data compression values I had hoped that it would "fit" into the "S0" pricing tier or "S1" worst case. As an aside my understanding is that I cannot use the "B" tier because my source data is > 16GB.

S0 provides a 10GB cache while the S2 cache is 25GB. However in both cases I got "Memory error: You have reached the maximum allowable memory allocation for your tier. Consider upgrading to a tier with more available memory" about two thirds of the way through processing. Even though it would be way outside my budget, I tried upgrading to the S3 tier (50 GB) but still got the same error. I have read muitiple sites/posts but can't find anything definitive on how SSAS on Azure uses memory . Any guidance on how it works and/or what I'm doing wrong.

解决方案

Azure SSAS is an in-memory database, so you need enough memory to store the data you want to analyse. It sounds like you have more data than is currently allowed in the tier you have purchased. As the message says, you can upgrade to a more expensive tier which will allow you to work with more data; you should also look at the data you are loading and delete any columns, because that will reduce your memory usage.

You may check out the memory usage from Azure Portal or via Azure Server properties.

Analysis Services provides metrics to help you monitor the performance and health of your servers. For example, monitor memory and CPU usage, number of client connections, and query resource consumption. Analysis Services uses the same monitoring framework as most other Azure services. To learn more, see Metrics in Microsoft Azure and New memory options for Analysis services.

For more details, you may refer SO thread which addressing similar issue.

Hope this helps.

这篇关于允许的最大内存分配所有层的例外情况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！