feat: update cors headers for cross domain request

[KristinAoki]

Description

The import page chunks the uploaded file and uses the network headers content-range and content-disposition to send the chunk to the backend. These headers are not part of the default cors headers. This change impacts Authors.

Supporting information

JIRA Ticket: TNL-11163

I was able to import a 60MB course. When I tried to do a 250MB course it broke. I tried a bit smaller and it seemed to break around 100MB. It looked like the percent was going up. The percent got to 99%, then the percentage disappeared and the gear would just spin, see screenshots. The legacy page they can import a 500 MB course but with the new page cannot import 100 MB.

Testing instructions

Testing

1. Navigate to the Import page
2. If you wan to keep the content of your current course, export it before importing a new one
3. Upload a course zip that is greater than 100 MB
4. File should import
5. Wait for the import to finish
6. Click "View updated outline"
7. Check that the course properly imported

Deadline

None

Other information

frontend-app-course-authoring PR #734 is dependent on this change.

[bszabo]

Can you please clarify how these configuration changes fixed the problem? Is it the case that for files under the 100MB threshold no chunking is required, and therefore the offending headers (relative to CORS) are never used? But that chunking occurs with files bigger than this, that the headers are then introduced, and that CORS then blocks the operation?

[KristinAoki]

Can you please clarify how these configuration changes fixed the problem? Is it the case that for files under the 100MB threshold no chunking is required, and therefore the offending headers (relative to CORS) are never used? But that for files bigger than this chunking occurs, the headers are introduced, and CORS blocks the operation?

Correct, files under 100MB do not require chunking because they are within the accepted file size range. However, when a file >= 100MB is submitted a 413 error is thrown. When no chunking is required, no additional headers are required on the request. However, the chunking requires the additional headers for the backend to know how to stitch the chunks back together. As a result, the additional CORS headers need to be added or the API will always fail with a CORS error. It should be noted, that regardless of file size all files will go through the chunking process when importing in the course authoring mfe.

[bszabo]

Per our slack huddle, where we observed this PRs changes importing a 2GB file on hosted devstack, import looked to be in 1.2 kByte chunks, each taking a few seconds. A back-of-the-envelope calculation suggests that on the order of one million chunks would be required to get to a 2GByte file. A 10 minute delay would be 600 seconds, or maybe 200 chunks. This suggests that the solution we observed would not complete in anything near an acceptable timeframe.

[bszabo]

Based on feedback that upload of 2 GB file took 2 minuts, lgtm

[edx-pipeline-bot]

2U Release Notice: This PR has been deployed to the edX staging environment in preparation for a release to production.

[edx-pipeline-bot]

2U Release Notice: This PR has been deployed to the edX production environment.

[edx-pipeline-bot]

2U Release Notice: This PR has been deployed to the edX production environment.

[edx-pipeline-bot]

2U Release Notice: This PR has been deployed to the edX staging environment in preparation for a release to production.

[edx-pipeline-bot]

2U Release Notice: This PR has been deployed to the edX production environment.