From scratch
1. run sp_plan_dbccdb from the master database to get an estimate of the sizes needed
2. create a dedicated data and log device for the dbccdb database and create the database onto these, or just create it on existing devices
3. run isql –U –P –S –i$SYBASE/$SYBASE_ASE/scripts/installdbccdb to install all the necessary tables and stored procs in the the dbccdb database
4. Add segments into the dbccdb database;
use dbccdb
go
sp_addsegment scanseg, dbccdb, dbccdb_data1
go
sp_addsegment textseg, dbccdb, dbccdb_data1
Where dbccdb_data1 is the data device dbccdb was created on
5. Add a 16K pool if it doesn’t already exist;
sp_poolconfig "default data cache", "150M", "16K"
go
5. Create workspaces;
use dbccdb
go
sp_dbcc_createws dbccdb, scanseg, scan_pubs2, “scan”, “10M”
go
sp_dbcc_createws dbccdb, textseg, text_pubs2, “text”, “10M”
go
The values in MB are found from the highest values of the output from sp_plan_dbccdb earlier
6. Configure databases into dbccdb; derive the required values from the sp_plan_dbccdb earlier
use dbccdb
go
sp_dbcc_updateconfig pubs2,”max worker processes”, “4”
go
sp_dbcc_updateconfig pubs2, “dbcc named cache”, “default data cache”, “10K”
go
sp_dbcc_updateconfig pubs2, “scan workspace”, scan_pubs2
go
sp_dbcc_updateconfig pubs2, “text workspace”, text_pubs2
go
sp_dbcc_updateconfig pubs2, “OAM count threshold”, “5”
go
sp_dbcc_updateconfig pubs2, “IO error abort”, “3”
go
sp_dbcc_updateconfig pubs2,”linkage error abort”, “8”
go
You will need to increase the number of worker processes parameter to 100 and bounce the Sybase server, after it comes up set max parallel degree to 10 and max scan parallel degree to 3
