DevOps and the DBA: The Whole Package
For as long as there have been databases, there have been database administrators. Due to the complexity of relational databases, these experts became focused on the care and management of their database platform as the field matured. With feature enhancement and the introduction of DevOps, the DBA is now being asked to branch out beyond the database tier. Where once most focused their skills on a few areas in a platform, we now grasp the need to utilize them more and DevOps introduces some real opportunities for Database Engineers in the way of:
- Release Management
With this change, DBAs are embracing the power of “packaging” their environment into multi-tier containers, using Docker, Kubernetes and other products. This is where data pods from the Delphix Data Platform are impressively powerful, not just to DevOps initiatives, but to what was day-to-day tasks. We’re fully aware that when a release occurs, we rarely, if ever just make changes to the database tier. The application, OS level files and data residing inside the database may have changed, along with mid-tier products, like SOA in Oracle and Power BI in SQL Server. With this kind of multi-tier interactivity, if something does go wrong, it requires a set of backout steps to “undo” the changes. If the development and test environments are packaged as data pods and have software managing packaging it and allowing it to manage the time flow of changes occurring, this can greatly ease the development and testing requirements involved to rollback if something unexpected occurs.
As we embrace and educate ourselves on more shell scripting skills, Python, Perl and automation tools like Ansible, Chef, Puppet and Terraform, we learn to appreciate the self-service features of some container products. A data pod through Delphix lets the developer or tester manage the pods changes to database, application and other tiers. They can create bookmarks to tag changes, refresh from the up-to-date source, (which pulls regularly from production) or rewind when something doesn’t go as planned. There’s no need for resources to be allocated to writing rollback scripts, saving DevOps considerable time and energy that’s a common strain in DevOps shops.
Many DBAs are familiar with data gravity: the propensity for applications and services to be built around data. We feel data gravity driving demand whenever a request comes in to build, refresh, or migrate data. But the natural life of a database is growing, and our tools aren’t keeping up with that growth - databases are still hard to manage, secure, and move. The resulting friction creates pain for DBAs every single day. The power of the Delphix Data Platform eliminates that friction for everyone connected to the data tier so that data can be as flexible and agile as the rest of the development cycle.
By removing friction and eliminating data gravity as a concern, the DBA can take on new challenges of further automation, scripting monitoring, and release management. The idea of this expanded role doesn’t seem so overwhelming and is actually welcomed.
DBAs, in turn, can become more competitive in the market. We become essential experts and a required member of the IT team and development cycle. Our tools become an extension of us, offering great value and freeing us up to take on the new technology challenges of tomorrow.