Fix Scheduler Index Rebuild

Steps to reproduce

The new scheduler has two problems which causes index rebuilds to not
work properly:

1. The index is rebuild only from objects which have the scheduler set
as owner in the asset manager. This is true only for upcoming
recordings but not for those which have already been processed,
causing technical metadata to not be present in the indexes which in
turn causes features like filtering for a capture agent identifier in
the admin interface to fail.

2. Updates are sent to all scheduler receivers. This means that if
someone rebuilds the admin interface Elasticsearch index, the
messages are sent to all other indexes as well and the data are
received by, processed and stored in the External API index as well.
When doing a full rebuild, in the end, this cases twice the amount of
messages to be sent for scheduler updates (essentially building both
indexes twice).

Fixed and reviewed


Lars Kiesow


Lars Kiesow


Data Loss/Corruption

Tags (folksonomy)



Affects versions