> create clustered index cdateidx on calls2(calldate)
> go
For some reason, I just don't know about this. Putting a clustered
index on the calls table is going to slow down inserts. Putting a
clustered index on the date is going to put stress on the last page
of the table, especially with SQL 6.5. This will cause deadlocks
if two RadiusNT boxes have active accounting streams coming in.
I typically do NOT put a clustered index on a table that is being
modified/updated a lot. There are better used for small lookup
tables where you are going to be pulling the data out in the
cluster order.
> alter table calls2 add constraint pk_calls2 unique (NasIdentifier, NasPort,
> AcctSessionID, AcctStatusType, UserName)
> go
....
> sp_rename calls2, calls
> go
....
> alter table calls add constraint pk_calls unique (NasIdentifier, NasPort,
> AcctSessionID, AcctStatusType, UserName)
> go
Why do this a second time?
--Dale E. Reed Jr. Emerald and RadiusNT__________________________________________IEA Software, Inc. www.iea-software.com
For more information about this list (including removal) go to:http://www.iea-software.com/support/maillists/liststart