Showing posts with label advise. Show all posts
Showing posts with label advise. Show all posts

Sunday, March 25, 2012

BEST RECOMMENDATION FOR A TESTING SCENARIO

guys:
I would like to have your best advise.
We am planning to install Passive/Active SQL server Clustering in the
production environment that includes two identical Dell PowerEdge 6650 dual
processor servers with a External Disk Storage PoweVault 220S. We have
bought two copies of Windows 2003 enterprise (for each one of the servers)
and one only copy of SQL Server 200 enterprise Edition licensed per
processor (I was told that only one copy of this software ins needed in an
Active/Passive mode)
On the other hand, since the configuration above would be in production at
all times, we would like to have similar scenario for development and
testing purposes.I was advised to replicate the same hardware and software
scenario described above, however, as you can see, it would be a costly
endeavor. Each Dell 6650 costs approximately $20,000 (two processors, 8 GB
RAM, 2 MB cache) the External Raid $8,000, the Windows 2003 Enterprise
server software $3,000 (for the two servers), and the SQL Server 200
enterprise software about $24,000 (license for two processor, for just one
server)
Could you please if you see any issue in the production scenario? Are we Ok
using only one copy of SQL server there?
Secondly, do we have any other choice for a development and testing
scenario? Most people recommend that a development and testing scenario
would be, if not identical, at least very similar to the production
scenario. I was planning to get those Dell 6650 server but with a single
processor, only 1 GB of RAM, and 1 MB cache (or even 512 KB). In terms of
software I was also told that one economical approach would be to acquire
the MSDN Universal subscription that allows use software for testing
(Windows and SQL server)
Thirdly, do you have any other (most economical recommendation in terms of
hardware and software for our development and testing scenario? How critical
is that this has to be very similar (or identical) to the production
scenario?
Thanks for all your answers
White
First off, what you are implementing is a single instance cluster, not an
active/passive cluster. It may just look like a term, but there is a very
dramatic difference between the two.
As for licensing SQL Server in a cluster, you have exactly what you need.
The easiest way to tally up licensing for a cluster is to ask how many SQL
Servers you can connect to from an application. In your case, it would be
1.
For the dev/testing environments, you do not need to purchase the Enterprise
Edition of SQL Server. You can use the Developer Edition which gives you
the full functionality of Enterprise Edition without all of the cost and
hardware requirements. This even allows you to simulate a cluster, you just
don't get full clustering functionality. But, you do NOT need to stuff
clusters into your dev/test environments. There is no case that I'm aware
of where a clustered SQL Server behaves differently with respect to an
application than a standalone SQL Server.
My recommendation would be to purchase a Dell 6650 with the external RAID
array. Depending on your testing and development scenario, you can very
easily place BOTH dev and test on the same machine in different SQL Server
instances without collisions. The only real reason to have completely
different systems for the two would be if you are doing a lot of very heavy
performance related work. If not, you can get away with a single machine
with external array + Windows 2003 Server + SQL Server 2000 Developer
Edition.
I can send you my address so you can send a check for the ~$54,000 that I
just saved you.
Mike
Principal Mentor
Solid Quality Learning
"More than just Training"
SQL Server MVP
http://www.solidqualitylearning.com
http://www.mssqlserver.com

Best program for SQL database manipulation

Hello All

I am a relative beginner to SQL databases & new to this forum, so please bear with me if my query is too basic and advise if this question belongs somewhere else

I began working at a company that uses a program that stores data in an SQL database running off a Firebird engine

The program itself doesnt come with database management/administration module, so I'll need to use an external program for manipulating data in the tables that the database contains.

I have relatively little knowledge in SQL programming, which is why I would like to know which is the most powerful program for SQL database updating / manipulation?

This database has tables that has an infinite number of joins with other tables - Even MS Access wasnt able to open a few tables in this database because of the number of joins. I have tried Access & Lotus Approach, Approach manages to do a better job than Access, it open the tables & seems like it will let me import external data directly into the SQL table, but takes forever & usually just bums out giving an error after a very long wait..

My question is - Apart from Access & Approach, are there any more powerful, yet user friendly programs out there that can help me update data directly into SQL tables? What options do I have - the tasks I need to perform are pretty simple updating & cleaning of data already in there

Please, I will hugely appreciate any pointers that you guys the experts might have for me in this regard

Thanks
AlexTry Microsoft SQL Server 2005 Express. It's Free & Downloadable From http://msdn.microsoft.com/vstudio/express/sql/sql

Sunday, March 11, 2012

Best Practice for Ranking

Looking for advise or best practice for ranking of top 10 companies (out of 100s) for aggregates on a shipping comparitive value (tonnage). Would like to make each of the top 10 companies a column with values broken down by region (rows). I also need a column for a single chosen company (not necessarily top 10), and another column for that company's percentage of market share.

I'd like to utilize our data cube in a matrix, but can't solve how to add the last two columns above. We have the option of using the cube or the transactional db. Any advice is helpful, thanks.

One option would be to use the sort number 1-10 as your grouping key and have your report dataset based on a union of the 1-10 dataset, the 11 extra column dataset, and 12 extra column dataset. Then do a Matrix to line it up.|||Thanks, we'll give this a try.