Hyrax Fedora 6 Working Group - Phase 2 Performance Testing

[Arran Griffith]

Hello!

The Hyrax Fedora 6 Working Group is happy to announce that we are ready for the next phase of  Performance Testing on https://f6.nurax.samvera.org/ and we need your help!

Phase 2 Performance Testing - The performance testing use cases we are trying in Phase 2 required the buildup of a collection of 5,000 works to allow testing to happen involving adding, removing, and edit works within a collection of that size. Implementing Bulkrax on f6.nurax was necessary to create that size of a collection. Additionally, there are two test cases that require participants to use Bulkrax directly to import, so we had some pieces to add to f6.nurax before this testing at this scale could occur. 

This testing setup will be available until August 31, 2026. 

• After this time, we will remove the collection with 5000+ works so as not to incur additional AWS costs due to site hosting so we are encouraging all those who are interested to do so before the 31st.

Now that we are ready, here are the steps to give this a try:

1. Visit the working group wiki if you need to sign up for an account on f6.nurax: https://samvera.atlassian.net/wiki/spaces/samvera/pages/3223289860/Community+Performance+Testing+-+f6.nurax
2. Community Performance Testing - F6 Nurax spreadsheet, tab #3 for Phase 2 (Bulkrax/5000+ works) - https://docs.google.com/spreadsheets/d/1EsWzLPdopXn_xRGlLc7OEaxxnsvTug2o0g5I1DNIttU/edit?gid=1401253569#gid=1401253569

The first 2 test cases involve a single work with over 150 files that you bulk ingest and then delete. There is a template Zip file attached to the working group wiki page for importing a work with 152 pages (zip contains 152PageWork.csv and “files” folder with .png image files) so you don't have to come up with this work yourself: Phase2-ZipImport-152Pages.zip

You can use the .csv file as it is within this .zip or update source_identifer, title, and creator values. Do not change model, file, or parents values. If you do update the .csv file, you will need to create a new .zip file containing just the .csv file and the “files” folder that contains the 152 page images. See Importing from a Zip file for more details. This Zip import should add 1 work on f6.nurax. The next test case is to delete this work so this should prevent any space issues on this testing site.

The next 3 test cases are working with the collection that contains over 5,000 works already: Bulkrax Test Collection 

There is another template Zip file attached to the working group wiki page for importing a work with 1 file (zip contains 1FileWork.csv and "files" folder with 1 .png image file): Phase2-ZipImport-1FileWork.zip

You can use that .csv file as it is within this .zip or make similar updates as described above. We recommend keeping this work small (only 1 file). This import will create this work. After the work is created, add the work to the Bulkrax Test Collection and make sure that succeeds. You can time both the import and adding the work to the collection but they are different activities.

We have one final test case that cannot be tested on f6.nurax since the rake task to run fixity checking via the user interface is not in place. It is an optional feature to include in a Hyrax installation and while it is something to consider for performance evaluation with new versions of Fedora, it is not as high of a priority to evaluate as the other test cases and we are forgoing that test case for now on f6.nurax.

On behalf of the  Hyrax Fedora 6 Working Group, we want to thank you for participating in this testing. Please reach out on #fcrepo on Samvera Slack if you have any questions.

Cheers,

Arran & Julie, on behalf of the Hyrax Fedora 6 Working Group

---

Arran Griffith

Program Manager, Fedora Program

Lyrasis

Book a time to meet with me

https://fedorarepository.org/