I've set the size of a database to 6GB - yet have 4GB memory allocated to
sql server.
Just focusing on this issue , is it more effective to have the size of the
db less then the physical RAM of the machine?

I recently read some article saying that fragmentation and page split is
bad, it gives some
test that shows that using guid key( more fragments) scheme gives only some
1/10

In perfmon the dbcc logical scan bytes/sec counter display, well, lot of
activity. Does anyone know what this counter measures?
I thought DBCC activity but there is none on the server.
Thanks,

I have a query that performs multiple LEFT OUTER JOINs (about 5 or so).  The
main table (on the furthest left-side of the LEFT OUTER JOINS) contains
19,000,000 rows.  The final estimate on the Estimated Query Execution Plan
is 19,000,000 rows.  So far, so good.

 In our OLTP system, I see a SPID being blocked by itself when I do sp_who2.
Why does this happen and what do you need to look at to resolve it ?
TIA
MO

We have an OLTP app that does insert/update/delete to couple of tables on sql
server 2k. There are some select queries as well against the same db. The
updates clash with the selects. When I look at sp_lock, I see these types of
locks

I'm running into an issue referencing views on partitioned tables if the view
contains the $partition function.
This works...
USE TheDB

When I create a none-unique clustered Index, the SQL-Server automaticly
definies a (hidden) unifier column. This is used (among other things) for
Key-Lookups. It is documented, that the SQL-Server checks on insert (or
update) if a duplicate exists, and increments the ...

One of our customers has strange problem with their SQL Server. We have
written an SP for them that retrieves 12  rows from 12 light queries and
inserts into a table variable. Finally selects all rows from table variable
and returns to the caller.