Open mitchclemens opened 7 years ago
See this comment, it seems possible the caches might not have updated in 2 seconds if you're having unreliability. Especially for a media item where the blob might take a bit to write to the DB before the caches fire.
Longer term it may be more reliable to move to a polling model where we place restoring items in a queue and test for their existence in the database every couple seconds until they do exist and can be reserialized.
Changed the classification of this. I see it as an enhancement over a bug, since Unicorn is not really designed for being used in an active Content Editing environment.
Using Unicorn 3.1.4 we have noticed when a content author deletes content and then restores it via the recycle bin the restored item is not always re-serialized.
This most commonly occurs with media items. It appears the size of the media item is not a factor into the reserialization bug.
Has this been corrected in an update since 3.1.4?