Drupal Migration: A University's Journey in Large-Scale Web Content Migration
Introduction
A public research university in Melbourne, Australia, renowned for its TAFE and higher education courses, has been consistently ranked high globally, including in the Times Higher Education rankings. The university faced a major challenge in migrating its extensive, complex Drupal website to the latest version. This case study outlines how our team assisted the university in successfully migrating about 2000 pages of unstructured content to a new structured website format.
Challenges
Transitioning the university's website from Drupal 7 to the latest version of Drupal posed a significant challenge, involving the migration of approximately 2000 pages with diverse content types including Courses, Events, Interest Areas, and more.
Content Diversity: The migration involved approximately 2000 pages with various content types, including Courses, Events, Interest Areas, News, and more.
Technical Complexity: Significant differences in field structures between Drupal 7 and the latest version of Drupal, adding complexity to the migration process.
Manual Testing Hurdles: The necessity for manual visual regression testing, a highly time-consuming task, which extended the project timeline.
Project Scope: Emphasizing the scale and diversity of data, highlighting the project's extensive nature.
Our Approach
Tool Used for Pages Migration Testing
For the Pages Migration Testing, we utilized Diffy, a tool specifically designed for verifying URLs, Breadcrumbs, and Menus. Diffy functions by taking screenshots and comparing them with previous versions of the same components, thereby identifying any visual discrepancies. As QA engineers, our role involved meticulously reviewing the list generated by Diffy to determine whether the observed differences were intentional or indicative of regression defects.
Manual Verification for Media
For media content, manual verification was essential. This process included cross-checking filenames, tags, titles, descriptions, and other relevant metadata to ensure accuracy and consistency
Test Data
URL, Menu, and Breadcrumbs Migration: Tested over 1000 URLs.
Media Migration: Included more than 100 images, documents, and videos.
Test Environment
Setup: Created a separate Drupal test environment, mirroring the production setup, for the page migration process.
Visual Validation Regression Testing
Tool Usage: Implemented Diffy for Page Migration testing.
Environment Comparison: Configured Diffy to compare D7 production with the latest Drupal test environment.
Process:
- Captured baseline screenshots using Diffy.
- Executed Diffy to capture new screenshots from the D9 test environment.
- Analyzed the output, highlighting differences from the baseline.
QA Engineer's Role:
- Reviewed each screenshot to determine if changes were deliberate or indicative of regression defects.
- Documented identified defects in JIRA.
Manual Validation for Media
Procedure:
- Individually compared each image in D7 with its counterpart in the new Drupal system implemented.
- Verified consistency in tags, descriptions, image formats, and filenames.
- Applied the same verification process for videos and documents.
Defect Documentation: Captured any identified defects in JIRA.
Result
Results Overview
Process Efficiency: The migration process was characterized by predictability, accuracy, and an iterative approach.
Defect Rate: Maintained within industry standards, demonstrating the effectiveness of our testing and validation methods.
Coverage and Cost-Effectiveness: Achieved wider coverage in testing while being cost-effective.
Migration Success Metrics
Page Migration:
- Total Pages Considered: 1284
- Success Rate: 100% (All 1284 pages were successfully migrated).
Image Migration:
- Total Images Considered: 3266
- Success Rate: 97.45% (Approximately 3182 images successfully migrated).
Document Migration:
- Total Documents Considered: 1184
- Success Rate: 97.80% (Approximately 1157 documents successfully migrated).
