Pageiolatch ??

Hello all,

I have an aggregation job that was taking about 2h on a 20 mil records database. Since some 5 milion new records were inserted 2 days ago, the job is now taking more than 10h. I updated statistics, recompiled all the stored procedures and defraged all the indexes. Still 10h. I deleted those 5 mil records, still 10h. The database file is not fragmented, as it is the only file on that drive. All the data is local, nothing is transferred over the network. Nothing else is running on that machine, just SQL Server. I can see the job having long waiting times, reason being PAGEIOLATCH_UP and PAGEIOLATCH_SH.

Is there something else I can do to bring the job back to 2h?

Thanks a lot! Any idea will be appreciated!Have you tried rebuilding the indexes? I have found that when SQL Server indexes get heavy insert/delete action without rebuilding, they can sometimes give unexpectly big performance issues. (Defraging them did not make a lot of difference)|||Yes, tried it ... no luck :-(

Thanks!

Originally posted by dbabren
Have you tried rebuilding the indexes? I have found that when SQL Server indexes get heavy insert/delete action without rebuilding, they can sometimes give unexpectly big performance issues. (Defraging them did not make a lot of difference)