Data chat with Kyle Hailey on how data virtualization with Delphix can aid Oracle DBAs in their day-to-day tasks. We talk about the basic issues facing Oracle DBAs and how specific use-cases can be improved dramatically.
47. Delphix Server 4.1
• Platforms
Vmware
Amazon EC2
• Databases
Oracle
SQL Server
Postgres
Sybase
• App Data
• Data Masking
• Replication
48. How can Delphix help DBAs?
• Scratch environments
Testing one-off patches, patchset updates (PSU’s), and critical-patch
upgrades (CPU’s)
• SQL tuning
When tuning a specific SQL statement, how can you effectively
test the impact of…
- adding or dropping an index?
- gathering CBO statistics a bit differently?
…without getting an act of Congress/Parliament?
Some bio, 30 years in IT, 25 years in Oracle, 20 years as a performance-tuning DBA and as on-off Apps SysAdmin, books, ACED, OakTable, joined Delphix last May. Why?
DevOps. Increasing the tempo of application development. Supporting agile development methods. Shorter delivery cycles and continuous delivery. It isn’t just people working harder, it needs some magic.
Between a rock and a hard place
Due to the constraints of building clone copy database environments one ends up in the “culture of no”
Where developers stop asking for a copy of a production database because the answer is “no”
If the developers need to debug an anomaly seen on production or if they need to write a custom module which requires a copy of production they know not to even ask and just give up.
Before we proceed, there are a few terms that will come up frequently that we should define.
Within our production environment, we will frequently refer to the “source host” or “source database.”
We will also refer frequently to the Delphix Server, and to something called a dSource.
Finally in our pre-production environments we will talk about the “target host” and “VDBs.”
[read the slide]
[read the slide]
[read the slide]
Note that no Oracle processes run on the Delphix server.
[read the slide]
The dSource is typically a fraction of the size of it’s production database, due to several storage optimizations made by Delphix.
[read the slide]
[read the slide]
I should emphasize the first sentence again: VDBs are fully functional databases. They look like databases, their data files appear to be the same size as they would be physically. They can be backed up. They can be restored to. To end users, and to DBAs alike, they should be indistinguishable from any other database whose data files and logs happen to be stored on NFS.
With Delphix, the refresh process changes.
The refresh is accomplished by a few clicks in the Delphix GUI, via a command line interface, or via an automatic policy.
Instead of requiring a full terabyte of storage for the development server, we now require a fraction of that to store a compressed copy of the production database within Delphix. This becomes the basis of the storage for our new Virtual Database (VDB) in the development environment, and is provided to the development host via NFS.
As the development users begin to change their virtual database, only the changed blocks need be stored back within the Delphix server.
To summarize, the storage requirements to make a single copy have dropped from “as large as production” to “a fraction of production,” and the provisioning process is simple. Even better, additional copies require only a few changed blocks to be stored, so the more copies we make, the better the space savings become.
DevOps. Increasing the tempo of application development. Supporting agile development methods. Shorter delivery cycles and continuous delivery. It isn’t just people working harder, it needs some magic.