March 20, 2007 at 3:53 am
March 20, 2007 at 5:12 pm
Leo,
Did you try using detach/attach method instead of backup/restore?
thanks,
Olga
March 20, 2007 at 5:55 pm
Hmm - 500GB database That'd be scary!
Certainly the backup of a DB of 5-10GB in size doesn't take very long at all. A detach, file copy and then reattach could be a feasible solution as well, although I haven't really thought about it too far so there could something I've missed. For that matter, the detach & attach could take longer!? Benchmarking will tell I suppose.
I am curious though - what do you do when you have a 500GB database? I imagine some people might do something so horribly horribly crude as physically swapping drives haha
The idea about keeping the DBs the same and merely updating views to point to the correct DBs or, if a single DB with table suffixes, the correct tables seems logical, although maintenance of the views must be rememebered if and when a new table is added.
March 21, 2007 at 1:11 am
I can't remember when I last worked with a DWH of 5-10 GB My last was 1 TB..
CDR data are major league space consumers...
March 6, 2008 at 12:37 pm
Embedded images http://www.sqlservercentral.com/articles/2901/DataWa1.jpg and http://www.sqlservercentral.com/articles/2901/DataWa2.jpg are 404 when I look at this article, or if I try to hit them directly.
Gary
March 6, 2008 at 6:03 pm
This is pretty well done.
I do have a suggestion for an alternative that uses only one db.
You could do the ETL into different tables instead of different databases. Then switch a set of views to point to the new tables from the old. This might let you run without killing any users.
Switching db's has other advantages though.
Regards,
Andy
Andrew Novick
March 8, 2008 at 10:09 pm
Good solution for small Database but not for terrabyte databases. Ours takes 3 days to load from backup. Using partitioning and file groups or snapshot is a better choice
March 9, 2008 at 9:10 am
I dunno... to me, killing everyone's connection seems to violate a 24x7 SLA. I've used the same method at the table level before without killing any connections with no ill effects. Takes about 65 milliseconds per table.
--Jeff Moden
Change is inevitable... Change for the better is not.
March 9, 2008 at 9:13 am
Gary E. Vernon (3/8/2008)
Good solution for small Database but not for terrabyte databases. Ours takes 3 days to load from backup. Using partitioning and file groups or snapshot is a better choice
Heh... wow, Gary... does the boss know that? 3 days is an awfully long time to be "out of business". I'm thinking you and yours need to figure out a way to do some parallel loads to cut the time down.
--Jeff Moden
Change is inevitable... Change for the better is not.
March 9, 2008 at 7:05 pm
We are using connection pool. There is no direct customer connections.
March 9, 2008 at 7:09 pm
Detach attach is the problem because if process failed detach the database is not exists. If rename failed you have the old database working
March 9, 2008 at 7:15 pm
noeld (3/19/2007)
A 10GB data warehouse is small enough to follow the proposed procedure.When your data warehouse reaches 200 to 500GB you will get into a TOTALLY different ball game.
Cheers,
When you have 500G you strategy has to be changed. You may check if schema level can be replaced or another set of objects, or some other solutions. It is not as scary as people thinking. We have 80-100G statistical databases but the strategy is different.
March 9, 2008 at 7:19 pm
Nice article.
[font="Times New Roman"]-- RBarryYoung[/font], [font="Times New Roman"] (302)375-0451[/font] blog: MovingSQL.com, Twitter: @RBarryYoung[font="Arial Black"]
Proactive Performance Solutions, Inc. [/font][font="Verdana"] "Performance is our middle name."[/font]
March 10, 2008 at 5:53 am
Leo Peysakhovich (3/9/2008)
noeld (3/19/2007)
A 10GB data warehouse is small enough to follow the proposed procedure.When your data warehouse reaches 200 to 500GB you will get into a TOTALLY different ball game.
Cheers,
When you have 500G you strategy has to be changed. You may check if schema level can be replaced or another set of objects, or some other solutions. It is not as scary as people thinking. We have 80-100G statistical databases but the strategy is different.
So, what strategy do you use?
--Jeff Moden
Change is inevitable... Change for the better is not.
March 26, 2008 at 1:10 pm
Picture links are broken; is there an alternative location/URL for these links? Can I PM and get them e-mailed to me?
Thanks,
James R.
~ Without obstacles, you cannot progress ~
http://sqln.blogspot.com/
Viewing 15 posts - 16 through 29 (of 29 total)
You must be logged in to reply to this topic. Login to reply