Azure Analysis Services最大允许的内存问题 [英] Azure Analysis Services maximum allowable memory issue

问题描述

我有一个最近经常遇到的问题.我正在使用Analysis Services标准层，即.拥有20GB的内存.但是有时当我处理多维数据集时，会出现以下错误:

I have an issue which I am encountering quite often lately. I am using Analysis services Standard tier ie. having 20GB of memory. However sometimes when I process the cube i get the following error:

异常:Microsoft.AnalysisServices.OperationException:失败 将修改保存到服务器.返回错误:您已经到达 您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.技术细节: RootActivityId:e024b2d1-ce22-4edd-9a21-68b6b288d7bb日期(UTC): 8/11/2017 11:48:20 AM'.在 Microsoft.AnalysisServices.Tabular.Model.SaveChanges(SaveOptions saveOptions) Microsoft.AnalysisServices.Tabular.Model.SaveChanges(SaveFlags 在以下位置的Submission#0.Run(saveFlags)中运行(TimerInfo myTimer，TraceWriter log) :行59堆栈跟踪:在 Microsoft.AnalysisServices.Tabular.Model.SaveChanges(SaveOptions saveOptions) Microsoft.AnalysisServices.Tabular.Model.SaveChanges(SaveFlags 在Submission#0.Run(saveFlags)处运行(TimerInfo myTimer，TraceWriter日志)"

"Exception: Microsoft.AnalysisServices.OperationException: Failed to save modifications to the server. Error returned: 'You have reached the maximum allowable memory allocation for your tier. Consider upgrading to a tier with more available memory. Technical Details: RootActivityId: e024b2d1-ce22-4edd-9a21-68b6b288d7bb Date (UTC): 8/11/2017 11:48:20 AM '. at Microsoft.AnalysisServices.Tabular.Model.SaveChanges(SaveOptions saveOptions) at Microsoft.AnalysisServices.Tabular.Model.SaveChanges(SaveFlags saveFlags) at Submission#0.Run(TimerInfo myTimer, TraceWriter log) in :line 59 Stack trace: at Microsoft.AnalysisServices.Tabular.Model.SaveChanges(SaveOptions saveOptions) at Microsoft.AnalysisServices.Tabular.Model.SaveChanges(SaveFlags saveFlags) at Submission#0.Run(TimerInfo myTimer, TraceWriter log) "

基本上，表格使用内存引擎中的xVelocity，因此所有数据都在内存中.那么，这是否意味着该内存包括数据库中的所有内存，以及表格模型中Calculated表所需的所有内存?如果使用分区来处理模型，是否可以解决此问题?

Basically the tabular uses the xVelocity in memory engine so all the data is in memory. So does this mean that the memory includes all the memory in the database and also all the memory required by the Calculated tables in the tabular model please? And if one uses partitions to process the model would it solve this issue?

推荐答案

模型中的数据全部在内存中.如果只有一个分区已经被处理，并且您在其上调用了full进程，则将需要两倍的内存.对数据进行分区并仅加载新的/更改的分区将避免您需要双倍的内存.

The data from the model is all in memory. If you have only one partition that is already processed and you call process full on it, you will need double the memory. Partitioning the data and only loading new/changed partitions would avoid you needing double the memory.

这篇关于Azure Analysis Services最大允许的内存问题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！