I need to contrive a high availability configuration for Analysis
Services without using clustering; basically, each component of the
system has to be duplicated at two geographically remote locations so
that a catastrophic site failure at one location won't stop the
operation of the other. Because of the characteristics of the network
links between the two sites, clustering isn't an option.

Assuming two Analysis Services servers (one at each location), each
will be getting an identical copy of the source data, arriving at the
staging SQL Servers on a continuous basis. As the source data is
identical, each Analysis Server's database of cubes should be too.
For performance and load-balancing purposes, client queries could
(and should) be serviced by both servers. Operationally, if one of
the Analysis Servers fails to process a set of source data correctly
(and hence is out of step with the other Analysis Server) I would
need to prevent the problem server from servicing client queries
until it is back in step with its partner.

I doubt this can be the only time such a requirement has arisen, so I
wondered if anyone would share any experiences or recommendations on
this scenario?

Cheers,
--
Use Reply-To address or replace deadspam with hotmail

You might want to look at:
http://www.microsoft.com/sql/evaluat...gOLAPsites.asp
While it does talk mostly about NLB clustering, there are sections on
keeping two servers in-sync, etc. you might find useful.
--
Dave Wickert [MSFT]
dwickert (AT) online (DOT) microsoft.com
Program Manager
BI Practices Team
SQL BI Product Unit (Analysis Services)
--
This posting is provided "AS IS" with no warranties, and confers no rights.

"Koan G" <koangee (AT) deadspam (DOT) com> wrote