DevOps efforts stumble over database obstacles, survey suggests

Application code changes may be moving along swimmingly, but database schema changes are another story.

DevOps and Agile methodologies may help speed up and sync up software release cycles, but there's something that tends to get in the way of even the tightest operations -- back-end databases. Eight in 10 managers in a recent survey say that deploying database code changes takes longer than application code changes, and thus gums up the process. 

keyboard-and-exclamation-points-photo-by-joe-mckendrick.jpg

Photo: Joe McKendrick

This is one of the takeaways of a survey of 307 IT managers from Dimensional Research, sponsored by Datical. While Datical, a provider of database release automation tools, has a stake in this data,.the results point to a significant obstacle to efforts to step up the pace of software releases. And the pressure is on -- most application teams (90%) report that they face pressure to release applications more quickly, and 43% are releasing applications on a daily or weekly basis, up from 35% in the survey from a year ago. 

The survey covered application developers, application release engineers, enterprise architects, and DevOps engineers. Of those, 92% reported it is difficult to accelerate the deployment of database schema changes in an effort to match the pace at which they deploy application code changes. 

When it comes to the data side of the equation, things have not improved since last year's survey, the study's authors report. In addition, companies with the most ambitious application release cycles suffered the biggest impact from slow, manual deployment of database schema changes. 

In most cases, application changes require database changes. More than half (57%) of all application changes requiring corresponding database changes, the survey shows. Among those with faster release requirements, close to half (46%) of those that release daily or weekly say it is extremely or very difficult to accelerate the database release process. This is noticeably higher than the 37% reported by teams with release cycles of two months or more. 

"The process of deploying database schema changes is very complex and demanding," the survey's authors point out. "It frequently involves spending hours, if not days, manually reviewing and validating every database change script and schema change within those scripts." More than nine in 10 respondents, 91%, say they have to re-work database changes multiple times before they are ready for production deployment.  This is unchanged from last year's results.

The top challenge participants cite is the lack of tools to automate the database deployment process (50%). This was followed closely by long database change review and approval cycles (49%) and having a very manual deployment process with many steps that can fail (48%). Other challenges mentioned include regulatory and compliance concerns, testing challenges, dependencies on other applications, semi-automated testing, and a mismatch between data for the pre-production and production environments requiring manual data manipulation language, the survey authors report..

The survey also shows that 84% of application stakeholders had serious production issues due to database change errors; 57% of all application changes require a corresponding database schema change; and 88% report taking more than an hour to resolve these schema change issues, up from 81% a year ago. 

The report highlights greater automation as the most effective way to address this lagging piece of the software release picture. When application stakeholders were asked about automating database deployments, nearly all (92%) say it would accelerate the overall application release cycle. Of course, DataOps is another emerging practice that will help address and automate the data delivery process. Stay tuned.