sql server full text search:
parkchangsoon,
The limited duration high CPU usage you are seeing during your scheduled
Incremental Population is expected and normal as there is a lot of internal
work being processed that is not only CPU intensive, but memory and disk i/o
intensive, but only for specific durations. You can reduce the priority of
the MSSearch service via sp_fulltext_service 'resource_usage' <value>, where
<value> is by default set to 3, and set it to 1 (background). However, this
will have the affect of causing the Incremental Population to take much
longer time to complete.
Since you are using SQL Server 2000, I'd highly recommend that you look at
"Change Tracking" and "Updated Index in Background" (see the BOL for more
info on Change Tracking) as this will eliminate the need for your scheduled
Incremental Population and allow your FT Catalog to be updated in near
real-time as the FT-enabled column is updated. This will eliminate the high
cpu usage you are seeing during the scheduled Incremental Population.
Regards,
John
[quoted text, click to view] "parkchangsoon" <dev_sr@albaroot.com> wrote in message
news:#PZ$lvLXEHA.2972@TK2MSFTNGP12.phx.gbl...
> os : win2000
> sql : sql2000 enterprise
>
> table name : job
> I created fulltext catalog
> table is job column name : title
>
> full population is every sunday am 5:00
> incremental population is every 2hours from 7am to next day 3am
>
> whenever server is running the both population, cpu is full 100%(see the
> task manager )
>
> Question : can i reduce the cpu use?
>
> the columns are updated frequently are readcnt of job table
>
> if i remove the readcnt column, the cpu use will reduce?
>
> thanks...
>
>
>
> below is the result of sp_spaceused job
>
> name rows reserved data index_size unused
> ------------------------------------------------------------
> Jobs 259845 468808 KB 214264 KB 28128 KB 226416 KB
>
> column_name type
> ------------------------
> title varchar(100)
> findPrsn tinyint
> ageZn varchar
> applyRght varchar
> emplrId varchar
> seq int
> readCnt int
> sido tinyint
> workTyp char
> applyDcmnt varchar
> findYmd2 smalldatetime
> gugun tinyint
> finishYmd bit
> ymd smalldatetime
> particular char
> position varchar
> logo varchar
> endymd smalldatetime
> jobtype1 tinyint
> jobtype2 tinyint
> jobtype3 tinyint
> paytype char
> pay1 int
> pay2 bit
> suprise tinyint
> ts timestamp
> visible bit
> createdYmd smalldatetime
>
>