Multitenant Snapshot Copy using ZFSSA [Thin Database Cloning]

Starting with Oracle 12.1.0.1 it was also possible using snapshot clones with pluggable databases within the multitenant database architecture. It was limited to file systems that support sparseness and ACFS. With the first patch set 12.1.0.2 it’s possible to use direct NFS as well for snapshot copies.

Multitenant thin cloning can be managed on Oracle storage like ASM:ACFS and ZFSSA. Special thought here is not to use the clonedbparameter. This will force to use the thin cloning on database level instead the snap and clone technics inside the ZFSSA. For ZFSSA the snapshot copy command needs a key store keystore with the password of the ZFSSA server to communicate. The communication from the Database site is done with the executable ZFSSAADM which will be delivered with the Oracle software tree of the RDBMS. See my blog Using on ZFSSA:ZFSSAADM - Copy On Write (COW) on ZFS Storage Appliance (ZFSSA)

In this blog I will demonstrate and describes the steps to enable thin provisioning and instantaneous cloning features of Oracle Database 12c when it is deployed on Oracle ZFS Storage Appliance.

Thin Cloning of PDBs is a Snap with Oracle Storage

Creating pluggable databases with and without OMF structure; [CDB=OMF storage structure]

For creating a pluggable database you have to consider how to create the pluggable database, and what are you current CDB settings. In this post i have played with and without OMF and difference file structures inside ONE CDB database. Mostly pluggable database will be created from the seed database. What is needed when you want the oracle files of the pluggable database on an other file structure storage. What will you use, and when should you use what. After the createing of a pluggable database, you need possible extra steps inside the pluggable database to keep your optimalisize storage environment. Think about the steps that are needed with a single none CDB, those steps would also be needed on a pluggable database by or after the creating steps. When you don't think about this, the change is that you get a spaghetti mixture of files inside your environment.

Lets play;