|
Home > Archive > MS SQL Server OLAP > November 2005 > Cube processing with minimal downtime
You are viewing an archived Text-only version of the thread.
To view this thread in it's original format and/or if you want to reply to
this thread please [click here]
| Author |
Cube processing with minimal downtime
|
|
|
| I have a complex multi-step cube processing routine which I apply to my
Processing cube. After processing I'd like to switch information in that cube
(move, update,…) to Reporting Live cube as fast as possible. What is the best
way to do that?
| |
| RJ Smith 2005-11-25, 8:23 pm |
| There's a nice whitepaper on how to do this in AS2000, along with how to use
load balancing to eliminate any downtime. For AS2K5, you can use the
synchronize command. You can also just do backup/restore, if that works in
your situation (partitions less than 2GB in 2000, and the data volumes are
reasonable, etc).
--RJ
SQL2000:
http://www.microsoft.com/technet/pr...n/olapsite.mspx
SQL2005: http://msdn2.microsoft.com/en-us/library/ms186658.aspx
"Tim" <Tim@discussions.microsoft.com> wrote in message
news:F7251BC2-201E-4054-9AC6- 9A969DC93257@microso
ft.com...
>I have a complex multi-step cube processing routine which I apply to my
> Processing cube. After processing I'd like to switch information in that
> cube
> (move, update,.) to Reporting Live cube as fast as possible. What is the
> best
> way to do that?
| |
| Dave Wickert [MSFT] 2005-11-25, 8:23 pm |
| How much hardware do you want to throw at the problem? You can archtiect so
it is virtually zero downtime. Here is one approach used in large sites.
http://www.microsoft.com/technet/pr...n/olapsite.mspx
This assumes that your reprocessing involves structural changes. If you are
just refreshing data (partitions), then you can just reprocess the
cubes/partitions. When you are processing, the new data is in a "shadow
folder" and users see the old data. Once processing completes, the
switchover happens seemlessly.
--
Dave Wickert [MSFT]
dwickert@online.microsoft.com
Program Manager
BI Systems Team
SQL BI Product Unit (Analysis Services)
--
This posting is provided "AS IS" with no warranties, and confers no rights.
"Tim" <Tim@discussions.microsoft.com> wrote in message
news:F7251BC2-201E-4054-9AC6- 9A969DC93257@microso
ft.com...
>I have a complex multi-step cube processing routine which I apply to my
> Processing cube. After processing I'd like to switch information in that
> cube
> (move, update,.) to Reporting Live cube as fast as possible. What is the
> best
> way to do that?
| |
|
| Thanks Dave,
In my scenario I cannot make the cube available for users immeadiatly after
processing because I have to finish updating data in an relational dtabases (
for instance copy some data from the cube). But after that I trying to switch
cube as fast as possible. Backup/Restore or synchronizing are solid but not
fast enough ways. I trying to Move Directories after Stop OLAP service -
fast , but not sure I do a right thing.
Please advice.
Thanks,
Tim.
"Dave Wickert [MSFT]" wrote:
> How much hardware do you want to throw at the problem? You can archtiect so
> it is virtually zero downtime. Here is one approach used in large sites.
> http://www.microsoft.com/technet/pr...n/olapsite.mspx
>
> This assumes that your reprocessing involves structural changes. If you are
> just refreshing data (partitions), then you can just reprocess the
> cubes/partitions. When you are processing, the new data is in a "shadow
> folder" and users see the old data. Once processing completes, the
> switchover happens seemlessly.
>
> --
> Dave Wickert [MSFT]
> dwickert@online.microsoft.com
> Program Manager
> BI Systems Team
> SQL BI Product Unit (Analysis Services)
> --
> This posting is provided "AS IS" with no warranties, and confers no rights.
>
>
> "Tim" <Tim@discussions.microsoft.com> wrote in message
> news:F7251BC2-201E-4054-9AC6- 9A969DC93257@microso
ft.com...
>
>
>
| |
| Dave Wickert [MSFT] 2005-11-29, 3:23 am |
| You cannot copy individual databases; you must copy the entire data folder.
If you look at the data folder root, you will see a file called master.vmp.
It contains the metadata version stamps for all of the AS objects in the
system.
If it gets mismatched with the objects in the database folder, then the
database objects are corrupted.
Thus while you *can* copy files between servers, you must copy the ENTIRE
data folder. Unfortunately that is a limitation of the system.
As you've discovered the supported mechanisms for moving data between
servers are the database (and server) sychronization and deployment wizards.
That is what they were designed to do.
--
Dave Wickert [MSFT]
dwickert@online.microsoft.com
Program Manager
BI Systems Team
SQL BI Product Unit (Analysis Services)
--
This posting is provided "AS IS" with no warranties, and confers no rights.
"Tim" <Tim@discussions.microsoft.com> wrote in message
news:A1157C4C-D77F-4DE9-989A- E1A0CB6C68A3@microso
ft.com...[color=darkred]
> Thanks Dave,
> In my scenario I cannot make the cube available for users immeadiatly
> after
> processing because I have to finish updating data in an relational
> dtabases (
> for instance copy some data from the cube). But after that I trying to
> switch
> cube as fast as possible. Backup/Restore or synchronizing are solid but
> not
> fast enough ways. I trying to Move Directories after Stop OLAP service -
> fast , but not sure I do a right thing.
> Please advice.
> Thanks,
> Tim.
>
> "Dave Wickert [MSFT]" wrote:
>
|
|
|
|
|