By alejandro.vargas on December 25, 2007 9:29 AM
Last week I had the privilege to share experiences about Oracle technology with a group of colleagues in Istanbul.
During the meeting we focused on Maximum Availability Architecture
topics, one of the subjects I did present was Database Migration to
ASM.
I think this is an important subject, many critical
production databases based on 8i and 9i are on the planning stage of
migration to 10g with ASM.
Rman provides a very efficient mean
to perform the migration, with full control on the hands of the DBA, a
minimum downtime window can be achieved, even for databases on the
multi-terabyte range.
This document: Database Migration to Oracle Automatic Storage Management contains a step by step description of this technique and how to implement it.
You can read TonguçYilmaz notes on the seminar on his excellent blog: H.Tonguç
YILMAZ Oracle Blog
A few comments about my visit to Istanbul:
Istanbul is a vibrant city, modern and ancient , plenty of colorful restaurants and typical stores, rich in ceramics and jewelry that reminds us of the greatness of the Ottoman Empire.
The Mosques in the fog, the ships crossing the Bosphorous in all directions, the youth filling Liberty Street and Taksim Square made a strong impression on my memory. The taste of real Turkish coffee in the company of my dear colleagues, in this cold Istanbul winter, is something I'll remember for a long time.
From left to right: Alejandro Vargas,
TonguçYilmaz and
Hüsnü Şensoy
By alejandro.vargas on December 28, 2007 10:52 AM
Another interesting subject we tested during the seminar was activating a physical standby database for test, reporting, etc.
Usually we are using storage snapshots for this purpose; we mount a storage generated copy of the production database on a report server, perform recovery and use it during the day. The process of resynchronizing the storage copy is repeated each night.
Using a physical standby for the same purpose is possible on 10g when flashback database is enabled on the standby. A guaranteed restore point is created to be able to re-activate and resynchronize the standby when the period of report/testing is over.
This is a powerful feature, worth knowing and using it.
You can find step by step implementation details on this document:
By alejandro.vargas on December 31, 2007 1:52 AM
One of the interesting questions I've got during the seminar was related to the possibility to migrate a terabyte database from a little endian machine to a big endian machine, using an ASM diskgroup to transfer the data.
Unfortunately ASM diskgroups are not cross platform enabled. There is no problem to mount the same diskgroup on multiple servers at the same time, provided that the OS system for all the servers is the same. You can also unmount the ASM disks from one server and mount them into another, but still it is required to have the same OS to be able to work with the diskgroup on the new server.
The alternative solution is to use Cross-Platform Transportable Tablespaces (XTTS). There are several possible strategies to accomplish the migration of a terabyte database using XTTS. On this post I'm presenting and example of how this technology is implemented, step by step
