Database Reference
In-Depth Information
environment, only one full copy is required. Additional environments are then created based on the clone, but with
independent data files. These data files contain only the changes made during testing. Experience shows that test
environments rarely change a large portion of the operational database. To benefit from the new method of cloning
databases you need to use character based storage: the network file system (NFS). This works hand in hand with many
organizations' desire to make use of cheaper storage tiers for non-critical workloads. You should not be fooled by the
lower price though, and perform testing to ensure that the new method of cloning database environments matches
your performance expectations. Figure
2-7
shows the minimum setup needed for using clonedb. The production
and development server are both connected to an NFS filer using dual 10GBit Ethernet. The NFS filer exports two
file systems. The export
/m/backups
is used for the production database image copy. The other export
/m/CLONE
will be used for the clone database's data files and is mounted to the development host. The infrastructure is likely
supplemental to an existing backup configuration but it is of course possible to take regular file system backups from
the NFS mount. You should check with the business users about their expectations regarding Recovery Time Objects
and other details around backups. Normally the hosting platform will use RMAN backups directly, the clonedb
infrastructure exists in addition to it.
Figure 2-7.
The minimum architectural components for copy-on-write cloning
Note the key feature of direct NFS here. Oracle's direct NFS client—available as part of the 11g release—is
entirely written in user mode, potentially allowing environments to benefit from less overhead compared to
kernel-NFS. Another great advantage is the availability of direct NFS on Windows, a platform that does not
otherwise benefit from NFS.
■
the use of the clonedb feature requires the use of direct NFs. the setup of directNFs is out of the scope of this
chapter but is available from the author's website.
Note
Overview
The process to be followed for the new feature is remarkably simple and comprises these steps:
1.
Create an image copy of your life database
2.
Make this backup available on the test server using direct NFS
Search WWH ::
Custom Search