using Oracle as the database for Ambari, Hive and Oozie, how do we go about sizing the each DB? also, which versions of oracle do we support 11g, 12g?

[#1]

We support both Oracle 11g r2 and Oracle 12c.

[#2]

Hi Martin,

According to our docs Oracle 11g R2 and 12c are the supported versions of oracle - http://docs.hortonworks.com/HDPDocuments/HDP2/HDP-2.3.2/bk_HDP_Reference_Guide/content/ch_supported_...

[#3 · Accepted solution]

@awatson@hortonworks.com

Yes, both versions are supported and thanks for sharing the doc link.

About sizing:

Ambari, Oozie , Hive - For small to large clusters, the database size and load wont be high unless there is something wrong with the setup.

Coming from Oracle DBA background, I would say that following configuration is good start

64 to 128 GB memory

16 to 32 cpu ( Dual core)

OS disk - standard size ( raid 10)

/oracle - 50 to 100 (raid 10) - oracle binaries

/oraclelogs - anything near to 500GB (raid 1, its for oracle redo log files , not archive logs)

/archivelogs - 300 to 500GB ( raid 1, oracle archive logs and make sure there is archive log rotation in place)

/data - 500G to 1 TB ( raid 10, oracle tablespaces)

Export backups MUST

Any good DBA will stick with both RMAN and Export backups.