Backup strategy for a Terbyte Data warehouse ?

[copernicus]

Hi,

I am working on a tera byte data warehouse project.I am not clear about this

a)Do I need to run the db in archivelog mode ?
b)But If I run in archivelog mode ,it will generate a lot of archives during the load,so do I put the tables in nologging mode ?
c) I intend to use RMAN backups.
Level 0 is complete backup.Is this the best way to go ?


Plz help.


-Copernicus

[skhanal]

a,b) If the only change that happens to data is during upload then use NOARCHIVELOG mode

c)Again if there are no changes, then take a complete backup immediately after the load and you are done.

[copernicus]

Thanks ..I will run the db in NOARCHIVELOG mode.

Plz comment on this ...

I want to put these INIT parameters.There will be potentially a large number of connections....

Are these parameters OK or do I need to modify anything.

-Copernicus


========================================
PGA_AGGREGATE_TARGET -> 1500 Mb
BUFFER_POOL_KEEP -> 500 Mb
BUFFER_POOL_RECYCLE -> 250 Mb
DB_CAHCE_SIZE in prod 1200 Mb
DB_CACHE_ADVICE -> ON
DB_FILE_MULTIBLOCK_READ_COUNT -> 32
HASH_JOIN_ENABLED -> TRUE
PARALLEL_MIN_SERVERS in prod -> 0
PARALLEL_MAX_SERVERS in prod -> 8
STAR_TRANSFORMATION_ENABLED -> true
QUERY_WRITE_ENABLED -> true
QUERY_REWRITE_INTEGRITY -> true
CREATE_BITMAP_AREA_SIZE----> 250 Mb
LARGE_POOL_SIZE -> 64 Mb

=====================================