#DivBlk: Principles in Action During a Website Migration
6) URLs and Scholarly Implications (Part II)
Semesters: Spring — Summer 2019
Committee: Co-chair Kelli Coles (PhD student), co-chair Michelle Byrnes (undergraduate), Lauren Cooper (librarian), Quader’a Henry (undergraduate) Caleb Trotter (CCP alum), and Keith Jones (Library IT Systems Programmer) with Jim Casey (co-director) providing strategic input
This is a good time to note another issue we were contending with: Our database file was enormous — roughly 5 GB due to spambots invading our Scripto transcription tool in 2018. The spambots added at least 200,000 pages of spam to our Mediawiki which threatened the security of not just the Library servers, but the entire UD campus network. This incident caused us to close down our transcription tool on the website indefinitely. Given the size of our database file, we were trying not to move the entire database to Reclaim Hosting; just our Convention documents.
We consulted with nearly everyone under the Omeka sun for advice and feedback, including UD’s Digital Scholarship and Publishing Team, UD Library IT, Princeton Center for Digital Humanities, University of Penn Price Lab and Library, Penn State’s Research Informatics and Publishing team, Reclaim Hosting, the Omeka forum, and Sharon Leon of Omeka.
Most everyone concurred this was an annoying problem and likely needed a script writer to help us with directing to the new Item ID URLs.
Another option we held was: What IF we moved the entire Omeka Classic installation in full to a test site on a clean install, and did the incremental updates that we were missing. Would we be able to keep our Item ID #s, and furthermore, could we prevent or fix any major issues that might arise? If all seemed well, we could then migrate this installation from test.coloredconventions.org to omeka.coloredconventions.org.
We posed this question, and it was Sharon Leon’s response of “That’s what I’d try” that confirmed our next game plan.
This is where the benefits of using a third party host laid bare the opportunity to test this theory out. We could stand up a test URL (literally, test.coloredconventions.org), and move the current CCP site wholesale to this server.
We set up two test instances on Reclaim Hosting to try two different methods. The one that worked was following the instructions at Omeka Classic User Files. [The other method we tried that did not work was showing an error, which may be related to the core code being modified.]
We migrated current Omeka 2.1 from the Library servers to a fresh, clean core Omeka 2.7 install on Reclaim. The Website Committee spot-checked 25 document items (13%) spanning the entire archive and confirmed ID #s were intact and the site appeared in working order. The one error we experienced was with viewing our Exhibits due to a code error. The Reclaim IT support team identified the issue and seemed to easily fix the problem. We then asked Reclaim to copy the test installation to the main installation at omeka.coloredconventions.org. Once again, we spot checked 30 Item IDs that remained intact. The Website Committee then split out the 180 Items to confirm the Item IDs were intact.
At the end of August, the Items were all made private in new Omeka and the team went through the items to only make public the necessary items. We were on our way!
With access to server space, a team of mixed but solid technical knowledge, and working with an IT company familiar with the platforms, we could test out theories. We could try. We could try and fail and learn from those failings. We could also push our technical capacities. Even if this solution had also failed, we could feel confident that we explored our options and did the best we could to maintain the undergird of our scholarly framework.
READ MORE: #DivBlk: Principles in Action During a Website Migration
The Colored Conventions Project was launched & cultivated at the University of Delaware from 2012-2020.