2.x branch creating duplicate media entities in Drupal

Chris Burge → made their first commit to this issue’s fork.

I opened an MR that also checks for queued assets. It's possible that an asset will be returned by multiple categories. Presently, that would result in the asset being added to the queue once for each category.

The bug described in the issue summary is caused by the media_acquiadam module. There's no issue on Drupal.org as Acquia uses Jira internally.

Summary is below:

Currently, the getAssetsByCategory() method of the Drupal\media_acquiadam\Client class doesn't use strict search matching. As a result, the query will return all categories with the word(s) in their name. For example, a search for "Collateral Catalog" will match both "Collateral Catalog" and "Collateral Catalog PDF".

That said, once that bug is fixed in the next release of media_acquiadam, we should still verify the behavior in this module if the same asset is in two categories.

The bug is in the queue worker which does zero validation to see if a queued item has already been imported by another process.

This bug has been latent for some time, but only surfaced by possible double-results due to non-strict searching be parallel processes.

This updated code will ensure queued jobs do not perform duplicated imports. The acquiadam_asset_import_cron could insert many jobs that are not executed within one cron process. The queue worker is set to process for 120, which should be more than enough time. But something could interrupt the process and leave jobs behind pending to be worked and double queued.

MR5 looks good. +1 for adding test coverage. MR4 failed to consider existing assets.

MR5 indirectly addresses the issue addressed by MR4 (by checking if a queued asset already exists before importing it). I still think that DamImporter->import() should keep track of what it's queuing so as to avoid putting duplicates into the queue.