Sizing Exadata - Oracle Exadata Recipes: A Problem-Solution Approach - page 119

Listing 5-16. lst05-16-archive-size.sql

SYS @ PROD> select avg(gbpd) AvgGBpd, max(gbpd) MaxGBpd,
avg(gbpd) * &&num_days AvgReq,
max(gbpd) * &&num_days MaxReq
from (
select al.dt, (al.blks * kc.blksize)/1024/1024/1024 gbpd
from
(select trunc(completion_time) dt, sum(blocks) blks
from v$archived_log
group by trunc(completion_time)) al,
(select max(lebsz) blksize
from x$kccle) kc
)
/
Enter value for num_days: 3
old 2: avg(gbpd) * &&num_days AvgReq,
new 2: avg(gbpd) * 3 AvgReq,
old 3: max(gbpd) * &&num_days MaxReq
new 3: max(gbpd) * 3 MaxReq

Avg GB per Day Max GB per Day GB Req Avg GB Req Max
------------------ -------------------- --------------- ----------------
6.11 123.16 18.34 369.48

1 row selected.

The output above shows that for this particular database, we need about 125 GB to store one day's worth of
archived redo on disk and approximately 370 GB to store three days on disk. Add the values for all of your databases
running in ARCHIVELOG mode to arrive at a total disk space requirement for Exadata. Typically, customers store
archived redo logs in a Fast Recovery Area ASM disk group. For details on this topic, please refer to Recipe 5-9.
Next, you should attempt to determine your current growth rates for your databases. We will cover this topic in
Recipe 5-6.

■ Note please reference recipe 5-6 to learn how to estimate your database growth rates using Oracle aWr data.

Hybrid Columnar Compression can offer a tremendous amount of disk space savings on Exadata. If you know
which tables, and their sizes, will be converted to Hybrid Columnar Compressed tables as part of the Exadata
migration, reduce your database size estimates by the appropriate sizing factor. We will not attempt to provide insight
on how to predict this in this recipe as mileage with Hybrid Columnar Compression will vary from organization to
organization and is predicated on how many and which tables are “fits” for HCC.

■ recipe 4-4 provides some insight on how to assess your workload and database for hybrid Columnar
Compression.

Note