It is a good idea to run multiple test migrations before proceeding to a production migration. It can help you identify and resolve potential issues with the content or with installation configurations.

At a minimum, test migrations of each different content type should be run to verify that each content type is handled as expected.

Objectives

• Based on the results of test migrations, determine whether data is ready for migration

Suggested Collaborators

• Project Lead
• Technical Lead
• Developers
• Functional Specialists, i.e. metadata librarians, digital preservation teams etc.
• SME on data and data migration

• IT Department for security and access requirements

Optional Activity

You can do a practice run by following along with this full migration demo, using a sample dataset provided the University of Wisconsin-Madison based on one of their collections found here. You can download the dataset in multiple zip files:

• datastreams 0-1.zip
• datastreams 2-3.zip
• datastreams 4-6.zip
• datastreams 7-9.zip
• datastreams a-b.zip
• datastreams c-e.zip
• datastreams f.zip
• objects.zip
• hakubra-llstore.xml

Instructions

1. Prepare data: if you have already exported and reviewed your data, and completed metadata remediation and mapping, then your data is ready for testing.
2. Select a set of PIDs that you want to use for testing.
3. Follow the instructions outlined in the Fedora 6.x Migration Documentation to complete test migrations and validate the results.
4. Debug and retest if necessary.

Result Analysis

1. How much time would be needed to migrate your entire repository?

2. Is the performance acceptable? Does the Fedora 6.x installation configuration need to be tuned to enhance performance?

3. What are the plans to minimize interruptions to users?

How will you implement quality control for the ingest of all repository collections?

This consideration should apply across your entire migration process - checking spreadsheets and other input before ingest (where relevant for the ingest method) and checking after ingest to ensure content appears correctly, all derivatives were generated correctly and that search is working properly across the site. See a talk from Amy Blau at Whitman College from IslandoraCon 2022 here:

** Here is a link to the metadata template and checklist mention in the talk: Whitman’s template & checklist for preparing spreadsheets.

Related articles

Related articles appear here based on the labels you select. Click to edit the macro and add or change labels.