May 26, 2015 at 11:11 am
Every other day or so I get a message from people in my company that cubes aren't working. Each time, I reprocess the dimensions and then process the cubes and they are up and running again. I make no changes. As I'm new to Analysis services, I'm not quite sure where to start looking for issues. Can someone point me in a direction to start troubleshooting?
Thanks.
May 26, 2015 at 11:33 am
Next time, before processing, check the status of the cubes and when they were last processed. That may shed some light on your problem.
For best practices on asking questions, please read the following article: Forum Etiquette: How to post data/code on a forum to get the best help[/url]
May 26, 2015 at 11:37 am
Alvin Ramard (5/26/2015)
Next time, before processing, check the status of the cubes and when they were last processed. That may shed some light on your problem.
Did that, there were data issues related to some keys. Data looks good though so my only guess was that something happened and then cleared up. Hard to diagnose though.
May 26, 2015 at 12:05 pm
JoshDBGuy (5/26/2015)
Alvin Ramard (5/26/2015)
Next time, before processing, check the status of the cubes and when they were last processed. That may shed some light on your problem.Did that, there were data issues related to some keys. Data looks good though so my only guess was that something happened and then cleared up. Hard to diagnose though.
From what you're saying, sounds like there might have been data issues at the normal processing time that were taken care of before the reprocessing time. I'd check to see if anyone, or any process, changed/updated the data between the processing time.
For best practices on asking questions, please read the following article: Forum Etiquette: How to post data/code on a forum to get the best help[/url]
May 29, 2015 at 1:35 am
Most probably it is the following reason:
You are reading from a OLTP database or a database that is updated constantly
June 4, 2015 at 1:10 pm
I'm a newbie at SSAS and I've been having a similar issue with dimensions failing sporadically. When we process a cube, there are some errors that occur when some of the dimensions are processed (i.e. "Process Dimension 'anonymous' failed' (I replaced the name of the dimension with anonymous). But then when we process the cube again, sometimes different dimensions receive errors even though no changes have been made to the structure of the cube, nor any changes to the data in the underlying OLTP database. It seems there is no rhyme or reason to these errors. One of our SSAS experts has pretty much verified that the structure of the cube/dimensions are sound. Might there be something corrupt in the database or server? Could this be due to a timeout issue or memory problem? Any help would be greatly appreciated.
Thanks!
June 4, 2015 at 2:31 pm
scott.marquardt 58097 (6/4/2015)
I'm a newbie at SSAS and I've been having a similar issue with dimensions failing sporadically. When we process a cube, there are some errors that occur when some of the dimensions are processed (i.e. "Process Dimension 'anonymous' failed' (I replaced the name of the dimension with anonymous). But then when we process the cube again, sometimes different dimensions receive errors even though no changes have been made to the structure of the cube, nor any changes to the data in the underlying OLTP database. It seems there is no rhyme or reason to these errors. One of our SSAS experts has pretty much verified that the structure of the cube/dimensions are sound. Might there be something corrupt in the database or server? Could this be due to a timeout issue or memory problem? Any help would be greatly appreciated.Thanks!
I think it's always a good idea to process the dimensions first followed by the measure groups; this way you avoid key errors in processing. However, an OLTP db as a source is far from ideal as the data can change in the course of a process. The data should be as static as possible when SSAS reads it. It might be worth checking your processing settings and making sure it is parellel rather than sequential as this may help matters. Additionally make sure if you are using views or named queries in the DSV that you are not using the WITH (NOLOCK) hint. Really though, you need to create a separation between your OLTP db and the data you use to process into a cube - even if it's just a simple staging environment.
June 5, 2015 at 12:38 pm
Thanks for the information and tips. When I tried processing each of the dimensions seperately, I discovered that certain dimensions were failing consistently. This helped me to pinpoint some flaws in the cube/dimensions. Thanks so much for your help!
Viewing 8 posts - 1 through 7 (of 7 total)
You must be logged in to reply to this topic. Login to reply