Xref: TK2MSFTNGP08.phx.gbl microsoft.public.sqlserver.clustering:14782
Hi,
We're in the process of architecting a very large SQL2K database. We
believe this database will grow 2TB per month. 99% of the inserts will be
done via bulk inserts. Approximately 3-4 per day an application that we
wrote will query and pull approximately 6 million records at a time.
* Will this work with SQL Enterprise or should I use DataCenter?
* Will this operate on Win2k Enterprise or should I use Datacenter?
* What I/O recommendations do you suggest (RAID or fiberchanner / RAID
array, SAN, or NetSAN, memory on controller, etc) to keep up with a bulk
insert of approximately 2,000 records per second?
* Will failover clustering work with a database this size (stupid
question but I have to ask)?
Thanks!
Jack
Comments Inline
Geoff N. Hiten
Microsoft SQL Server MVP
Senior Database Administrator
Careerbuilder.com
I support the Professional Association for SQL Server
www.sqlpass.org
"Jack Victor" <Nomail@.JustPostReply.tv> wrote in message
news:pan.2004.05.21.14.06.17.720000@.JustPostReply. tv...
> Hi,
> We're in the process of architecting a very large SQL2K database. We
> believe this database will grow 2TB per month. 99% of the inserts will be
> done via bulk inserts. Approximately 3-4 per day an application that we
> wrote will query and pull approximately 6 million records at a time.
> * Will this work with SQL Enterprise or should I use DataCenter?
DataCenter is an OS-level difference. SQL stops at Enterprise Edition, but
it will run on DataCenter OS. You may need DataCenter, but only for the #of
CPUs it will support.
> * Will this operate on Win2k Enterprise or should I use Datacenter?
>
Again, you may need the memory and/or CPU scalability of Datacenter. A lot
depends on the number of concurrent users and the total load.
> * What I/O recommendations do you suggest (RAID or fiberchanner / RAID
> array, SAN, or NetSAN, memory on controller, etc) to keep up with a bulk
> insert of approximately 2,000 records per second?
SAN. Big SAN with multiple HBAs and paths.
> * Will failover clustering work with a database this size (stupid
> question but I have to ask)?
Sure. It may take some time to go through the recovery process during a
failover, but clustering does not affect scalability.
> Thanks!
> Jack
>
Subscribe to:
Post Comments (Atom)
No comments:
Post a Comment