Re: SQL consolidation and user logins

Brian Lube ( brian@mpinet.net )
Mon, 10 Aug 1998 14:46:23 -0400

It sounds like the query is hurting you. Are you using a datediff to clear
it out (delete from calls where datediff(mm, calldate, getdate()) > 2)? If
you are this seems to be taking too long each night. I would suggest a
weekly our monthly removal (I'm using weekly now, and it only takes about
10 minutes each 7-10 days).

Brian Lube
MPInet

At 02:14 PM 8/10/98 -0400, you wrote:
>We are using RadiusNT 2.2 with ODBC and a MS-SQL 6.5 sp4 SQL database. We
>keep 2 months accounting for our records. Our database is generally has
>around 300MB to 400MB of space used to handle this. Every morning at 4AM
>SQL executive runs a script that deletes all calls records over 2 months
>old. According to SQL executive, this process has been taking between 12
>and 15 minutes to complete each night. Our problem is that during the
>consolidation process, there are periods of time in which no users can log
>onto the network. The Radius screen itself just freezes on an accounting
>request until consolidation completes. Originally Radius was installed on
>the same machine as SQL Server, but because of the login problems we moved
>it onto a different machine from the SQL server... no luck, still freezes
>during consolidation. The same copy of Radius is performing authentication
>and accounting. I suspect that the Calls table is getting locked during
>the mass delete query, but am not sure if this is causing our problem or
>not. Has anyone dealt with this? Is there a workaround that will allow us
>to perform consolidation without locking out logins? Would running two
>copies of radius on two seperate machines (one for authentication and one
>for accounting) help or am I heading the wrong direction with this?
>
>--
>
>
>
>===================================================
>= /\ Mike A. Miller =
>= /--\ \/ Abraxis Networks =
>=/ \ B R A /\ I S =
>= / =
>= N E T W O R K S michael@abraxis.com=
>===================================================
>