Fixes #29321 - Indexing large repo #8605
Conversation
|
Issues: #29321 |
sjha4
changed the title
|
Do you think it'd be possible to add a test that would have caught this (without having to actually sync a repo with thousands of units)? I'm thinking you could create 50,000 fake file units with ActiveRecordImport fairly quickly, and then grab all the ids and pass that to the sync_repository_association method? |
|
@jlsherrill : Added a couple of tests for the 2 types of flows, file and errata with 70000 units. Is that what you had in mind? |
test/models/file_unit_test.rb
Outdated
| ids_href_map["href#{i}"] = nil | ||
| i += 1 | ||
| end | ||
| Katello::FileUnit.import([:pulp_id], ids) |
There was a problem hiding this comment.
Adding validate: false shaves off about 20 seconds from this method call. It still takes about ~14 seconds, but i think thats okay. I don't know that i see a ton of value for having an errata test and a file test, any reason to have the errata test in addition?
There was a problem hiding this comment.
Right..I dont think we need the validate:true for this.
The errata test follows a different path where there are actually mappings in the id_href_map..So the 2 flows are the map having nil values for all keys on all content types minus errata and map having actual pulp3_hrefs for errata that we want to store on the katello_repository_erratum table.
|
Thanks @jlsherrill ! Merging.. |
To do: