We have recently upgraded some large databases (250GB +) from SQL 7 to SQL 2000.
The new SQL Server 2000 environment is installed on a much improved server, SAN and disk configuration (so we told). The server and all databases on the SQL 2000 installation were set to the collation of SQL_Latin1_General_CP1_CI_AS (I do not know if this is relevant to my query).
Our databases stored on a RAID5 disk configuration whereas the transaction logs are stored on a RAID1 (mirrored) disks. We have allocated 7GB out of 8GB of memory to SQL server.
Although most queries and READ operations are performing as expected (i.e. better than the old SQL7 environment) there is significant deterioration in WRITE operations such as UPDATE and rebuild of indexes.
For example, a clustered index 30 million rows table may now take 25 min to rebuild. It used to take 10 min. We can also see that the transaction log is growing in a much rapid pace (?) that it used to for similar operations.
Did anyone else experience a similar problem? Do we need to focus on SQL Server software or on the hardware/disks configuration?