高内存负载行为和数据卸载 [英] High memory load behaviour and data unloading

问题描述

你好，

我正在使用一个由三台服务器组成的集群，每台服务器都是Windows 2008 SP2 x64 Standard，1x Core 2 Duo @ 1,86GHz ，4GB内存，2x 80GB SATA硬盘，2x千兆以太网。
高速缓存中对象的大小通常为4 kB，我们的目标是尽可能多地存储它们（1000万或更多（在比我们现在更大的集群中，当然）） 
保持它们很长一段时间（几个月甚至更长）。 Velocity是否合适？
此缓存创建为NonExExrable。

第一个问题：
考虑到50％的RAM配额可用于数据缓存，应该可以存储多达6场演出，但我无法在那里演出超过5.3演出。怎么来的？

第二个问题：我写了一个生成数据并将其输入缓存的工具 - 当数据被集中放入缓存时，缓存有时会释放项目  （约5％） 
即使有足够的可用空间（当缓存中有超过1千兆的数据时，也会出现这种情况）。
为什么会出现这种情况？  Velicity监视器是否会提前加载并预先丢弃数据？

第三个问题：我用缓和项目填充缓存（即4在我周围玩耍的那些东西仍然有100万件物品。第二天早上我发现Velicity已经释放了一些（大约5％），即使没有获取/放置活动。
为什么会发生这种情况？

感谢您的回答

Lukas

解决方案

如果数据放在不可过期的缓存中，Velocity可以保存无限期的数据。

第一个问题：<系统上的可用内存可能已经超过总RAM大小的15％而且速度可能已经驱逐了对象，这就是为什么你不能放置超过5.3 Gigs的原因。
第二个问题：
行为来自GC，因为在操作过程中会产生一些临时对象，这些临时对象由GC间歇性地收集并释放存储器。

第三个问题：
是否有任何其他应用程序在该机器上运行，它正在获取RAM？

谢谢，

Anshul

&NBSP;

Hello,

I'm using a cluster consisting of three servers each of them Windows 2008 SP2 x64 Standard, 1x Core 2 Duo @ 1,86GHz, 4GB RAM, 2x 80GB SATA HDD, 2x Gigabit Ethernet.
The size of objects in cache is typically 4 kB, and our goal is to store as much of them as possible (10 million or more(In A bigger cluster than we have now, of  course)) 
and keep them for long periods OF time (months or even longer). Is Velocity suitable ?
This cache is created as NonExpirable.

the first problem:
Considering THAT 50% quota of RAM is available for data in cache, it should be possible to store up to 6 gigs, but I wasn't able to put there more than 5.3 gigs. How comes ?

Second trouble:
I wrote a tool generating data and entering these into cache - when the data are put into cache intensively, the cache sometimes releases items  (approx 5%) 
even if there is enough space available (this has occured when slightly over 1 giga of data was in the cache).
Why does this occur? Do velicity monitors load and preventively drop data in advance ?

Third problem:
I filled the cache with milion items (i.e. 4 gigs of memory), played around and there were still 1 million of items.
THE next morning I found out that Velicity has released some of them (again approximately 5%), even if there was no get/put activity.
Why has this happened?

Thanks for Your answer

Lukas

解决方案

The Velocity can hold the data with infinite expiry if the data is placed in the non expirable cache.


First problem:
The available memory on the system might have gone below 15% of the total RAM size and velocity might have evicted the objects, that is why you are not able to put more than 5.3 Gigs.

Second problem:
The behaviour comes from GC as during the operation serveral temporary objects rae created which are collected intermittently by GC and the memory is released.

Third problem:
Is there any other application running on that machine, which is acquring RAM?

Thanks,

Anshul

 

这篇关于高内存负载行为和数据卸载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！