Bulk Content Migration via API yields 23 failures - then 0 failures (process being done in AWS)
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
07-29-2020
02:23 PM
Due to some of the recent API changes noted on the "Canvas Platform Breaking Changes" Page, we must move several of our current processes to AWS. We are currently testing the process that populates all of our empty course shells each term. This process uses the "Content Migrations API / Create a Content Migration end point" to copy content into all of our upcoming sections, from a "Canvas Master Template" specific for each course title. (We do not use Blueprint courses.) Here's the link for the API we're using:
We're currently testing this out in both our Canvas Test and Canvas Beta environments. We've run into some concerns, and have some questions.
Yesterday, we ran the process in our Canvas Test Environment. We ran 800 migrations, and 23 of these failed. The reason for failure for all of these was "Error running course export". We decided to try re-running the process on Canvas Test again today, and there were 0 failures. We have 2 questions about this:
1. What types of things would cause these failures? We have had these errors in the past, but never more than 1 or 2 for around 1000 courses.
2. Since we are testing this in AWS, and this is a change, is there anything about this change that would be an issue? The AWS process is multi-threaded. Would that have any implications?
3. Is it ok to do multiple content imports for the same section? Is there anything that will cause an issue? It appears that the content is simply overwritten with each subsequent import (vs duplication or something else)? Are there any implications of multiple API content imports into the same course.
Thanks for any feedback!
Kathleen Barringer
Excelsior College
Instructional Systems Engineer
