Hi all,
I'm writing from Italy, sorry for my English...I try to do my best, anyway I got a little question about performance and concurrency with bigs tables.
My SQLServer have 4 CPU and 3 fisical drive over C:
- "E" stand for data file (.mdf)
- "F" stand for log file (.ldf)
- "G" stand for back-up files (.bak)
Now with got problems of lock on some big table used from customer care...have millions record.
Lock when comes are always 3-4 second and in many situation user experience, internal user, is bad 'cause there isslowdowns.
Now we have in plan to move this "X" table on partition schema or using filegroups.
I've reads documentation about conceptual of two possible implementations.
First of all is best have 2-3 new fisical disk, ok it's clear, and partitioning is more indicated for ETLapproach.
Filegroups are more simple to managment but the core of all is ndf files...best over disk.
For closing the question is:
Having no disk in addition a partition function can be a best solution for lock problems intheoretical line?
Havingnew discsI would beaddressed to use filegroup mixing data e index storing.
Thanks inadvance for the answers, ALEN