Rename big cluster results in outage of nextcloud

[sirdrakeistda]

Expected behaviour

Renaming big cluster should work without problems.

Actual behaviour

Ttrying to name a big cluster results in an outage of the nextcloud.

Steps to reproduce

1.Oupload many images of the same person (This could happen on a big event like birthday, party, wedding, ... were many images are created) 2.wait until facrecognition creates clusters of the person with >1000 images 3.Open facerecognition in Nextcloud and try to name the clusters -> loading the naming window takes very long and could result in a complete outage of the nextcloud until a restart

I'm not sure what facerecognition does in the background, so don't take my solution ideas too serious. For me it seems like all images are loaded in the background, even tough they are not displayed. Two things would be nice to solve this problem:

• optimize performance by reducing load in background (e.g. don't load images which are not displayed)
• add background command for naming clusters if the web interface performance is too bad

Server configuration

• Operating system: Debian bookworm

• PHP version: 8.1

• Web server: apache2

• Database: mysql

• Nextcloud version: 29.0.0

  Client configuration

• Browser: Firefox

• Operating system: Ubuntu

[sirdrakeistda]

For all people who face this issue: A workaround is the name person cluster option in memories (not merge) instead of the naming option in facrecognition itself. This option has a much better performance and wont crash the server.