Evaluating Oracle 11g Under Linux
[Posted By jdoor0 in DB Technology > Oracle]
I’ve been evaluating 11g running under Linux (Ubuntu) as a Windows port was not available the first few months. I find the automation of the day-to-day administration was significantly enhanced over what was present in 10g. The vision of installing the software, creating and dimensioning the physical data area (tablespaces, tables, indexes,…) and then from then on just let the database take care of the memory layout and occasionally run some jobs to enhance performance (”analyze”) has come quite close.
In fact, I find that the main job of a DBA has shifted to making sure the “secondary issues” are addressed (total disk/memory space of the system, storing backups/archives,…) – assuming the “logical” layout of the DB is OK. By the latter I mean “issues” that are hard/impossible to fix by just throwing more memory/CPU/disk at.
The SQL parse/execute engine has made significant advances (again), but it still cannot fix a wrong data/table layout (missing indexes, unnormalized tables,…) or clumsily written SQL statements, but it CAN at least warn you when things go awry…