Comments on: SSAS Weird Science https://thomaslarock.com/2009/05/ssas-weird-science/ Thomas LaRock is an author, speaker, data expert, and SQLRockstar. He helps people connect, learn, and share. Along the way he solves data problems, too. Wed, 06 May 2009 15:15:06 +0000 hourly 1 https://wordpress.org/?v=6.7.1 By: SQLBatman https://thomaslarock.com/2009/05/ssas-weird-science/#comment-573 Wed, 06 May 2009 15:15:06 +0000 http://sqlbatman.com/?p=1763#comment-573 In reply to Brandon Reno.

brandon,

good points all around, i was confusing the roles and functions of the services. however, i still think msft could provide a scaled down version of the agent with perhaps a sql express backend.

]]> By: Brandon Reno https://thomaslarock.com/2009/05/ssas-weird-science/#comment-572 Wed, 06 May 2009 03:15:24 +0000 http://sqlbatman.com/?p=1763#comment-572 SQL Server Agent uses msdb to store the jobs. what more would you expect?

where exactly are you going to ETL your data to without a database?

while it seems clever to have SSAS installed without the database components, I’m not sure it’s particularly useful.

]]> By: Steve https://thomaslarock.com/2009/05/ssas-weird-science/#comment-571 Wed, 06 May 2009 02:08:10 +0000 http://sqlbatman.com/?p=1763#comment-571 You could do what Cindy says, or just use Task Scheduler in Windows somehow? Call the dtsx from cmd line.

Wondering, is Profiler installed or not when you don’t install the engine? Can you profile SSAS?

]]> By: SQLBatman https://thomaslarock.com/2009/05/ssas-weird-science/#comment-570 Wed, 06 May 2009 01:20:47 +0000 http://sqlbatman.com/?p=1763#comment-570 In reply to Mike Walsh.

and i am not only talking about dumps, but what about the ETL process? if i scale out to an additional server, how would the ETL pieces be schduled?

]]> By: Mike Walsh https://thomaslarock.com/2009/05/ssas-weird-science/#comment-569 Wed, 06 May 2009 01:16:14 +0000 http://sqlbatman.com/?p=1763#comment-569 Great point. So you can do the approach Cindy uses, you can also use a 3rd party backup program to take a VSS snapshot of all of the files used in SSAS. Or you can rely on your definitions like we do. Basically we backup our cube definitions and disaster recovery plans call for restore of the relational sources but redeploy/process of the cubes. Will it take a bit longer? Perhaps but it meets our SLAs for the reports just fine.

I agree that a scheduler is an important missing piece. I also miss the ability to easily see what is slowing you down, what is causing your performance issues and react as it happens like you can in the DB engine. Yes you can run profiler, and you can make some config changes but the level of functionality brought on by the DMVs (or heck even SP_Who2 and KILL) is sorely lacking.

]]> By: SQLBatman https://thomaslarock.com/2009/05/ssas-weird-science/#comment-568 Wed, 06 May 2009 01:04:08 +0000 http://sqlbatman.com/?p=1763#comment-568 In reply to Cindy Gross.

thanks for the link Cindy. i still don’t feel that is an acceptable solution, because then I am relying on one agent to do the dumps for two instances, and i would prefer to keep everything local. if Denny Cherry can build a stand-alone agent for SQL Express, then Microsoft could give us the option to install SQL Agent when doing an install of SSAS without the db engine selected.

]]> By: Cindy Gross https://thomaslarock.com/2009/05/ssas-weird-science/#comment-567 Wed, 06 May 2009 00:43:33 +0000 http://sqlbatman.com/?p=1763#comment-567 You can use SQL Agent on another box to issue the backup commands to Analysis Services.

Backup Strategies with SQL Server 2005 Analysis Services
http://technet.microsoft.com/en-us/library/cc917611.aspx

]]>