企业绩效管理网

 找回密码
 立即注册

QQ登录

只需一步,快速开始

查看: 528|回复: 4

Same cubes, but different memory usage

[复制链接]

79

主题

383

帖子

562

积分

高级会员

Rank: 4

积分
562
QQ
发表于 2014-3-18 03:01:53 | 显示全部楼层 |阅读模式
Hi all,
I have 2 cubes on two different TM1 instances on a same server.
To me they are identical cubes. The data is the same on both cubes. Same number and order of dimension.

However, one of the cubes takes much more memory when loading and much more time to update.
Logging is inactive on both instances.
The parameters on both Tm1s are the same.

Any idea on where to investigate where difference lies?

Thanks
回复

使用道具 举报

78

主题

397

帖子

582

积分

高级会员

Rank: 4

积分
582
QQ
发表于 2014-3-18 04:49:58 | 显示全部楼层
Has one of the cubes been "optimized" that is has the internal dimension order been changed?
回复 支持 反对

使用道具 举报

83

主题

421

帖子

617

积分

高级会员

Rank: 4

积分
617
QQ
发表于 2014-3-18 05:42:35 | 显示全部楼层
jaudouig wrote:I have 2 cubes on two different TM1 instances on a same server.
To me they are identical cubes. The data is the same on both cubes. Same number and order of dimension.

You know (as opposed to assume) that they are identical because....? What I'm getting at is, did you copy exactly the same .cub file from one data directory to another? Come to that, are the two .cub files EXACTLY the same size?

(I was also going to ask whether one of the cubes had had its dimensions re-ordered but Lotsaram beat me to that one. That's a highly relevant question as well.)
jaudouig wrote:However, one of the cubes takes much more memory when loading

OK, but you are getting the memory usage for the cube (as opposed to the server session within Windows, which is a completely different matter) from where? Also, is Performance Monitor turned on in both sessions? If so, what does the breakdown of memory usage for each cube show in the }StatsByCube cube after startup?
jaudouig wrote:and much more time to update.
Logging is inactive on both instances.

Logging shouldn't have anything to do with load time unless a server goes down without a data save and has to reload a cube's data from the log file. Persistent feeders might though. Which means that:
jaudouig wrote:The parameters on both Tm1s are the same.

Have you confirmed this using a tool like , or just eyeballed it? Used to be in the olden days when .cfg files had but a few lines it would be safe to do the latter. These days with the files tending toward obesity it's really easy to miss a T in one place that is an F in the other so it's best to do an automated check.

本帖子中包含更多资源

您需要 登录 才可以下载或查看,没有帐号?立即注册

x
回复 支持 反对

使用道具 举报

64

主题

354

帖子

512

积分

高级会员

Rank: 4

积分
512
QQ
发表于 2014-3-18 06:47:12 | 显示全部楼层
The answer, my dear Watson, is that the two cubes cannot be identical. You are just currently unable to see in which ways they differ.
回复 支持 反对

使用道具 举报

76

主题

356

帖子

534

积分

高级会员

Rank: 4

积分
534
QQ
发表于 2014-3-18 07:31:57 | 显示全部楼层
I shouldn't have to ask this but does the cube have a rule, and if so does it make reference to other cubes? This is relevant whether the references are in the calcs section or in the feeders section.

Are the dimensions identical in the two servers, and are the attributes of the dimensions identical?

If there is a rule with external references then is the cube load order, and the feeder processing order the same?
回复 支持 反对

使用道具 举报

您需要登录后才可以回帖 登录 | 立即注册

本版积分规则

QQ|手机版|小黑屋|企业绩效管理网 ( 京ICP备14007298号   

GMT+8, 2020-9-20 06:50 , Processed in 0.192641 second(s), 13 queries , Memcache On.

Powered by Discuz! X3.1 Licensed

© 2001-2013 Comsenz Inc.

快速回复 返回顶部 返回列表