Delete all metadata of a CSW harvester

[rime1014]

Describe the bug
With GeoNetwork 4.2.5, we tried to delete all metadata for a harvester (>200000 metadata records) and encountered two problems:

1. Deleting all metadata of this harvester at once results in a system crash.
2. A discrepancy between the index and database was caused by deleting a large amount of metadata (deletion in steps of 1000).

To Reproduce
Steps to reproduce the behaviour of problem 1):

1. Go to admin console > harvesting
2. Click on the harvester
3. Select the tab 'X metadata records'
4. Scroll down to the list of all harvested metadata records of the selected harvester
5. Click on the delete button

Due to the large amount of metadata the harvester manages, the previous steps led to a system crash. It was not possible to delete all metadata of the harvester with the delete button.

Steps to reproduce the behaviour of problem 2):
As we could not delete all metadata simultaneously, we tried to delete it in steps of 1000 records.

1. Go to the 'new metadata' > edit View
   
2. Search for data assigned to the harvester
3. Deleting the data (several delete commands for every 1000 data records)
   
4. All data is deleted (search shows no data for the harvester)
5. Rebuild the index

The search displays data from the harvester again. The metadata was not deleted correctly from the database.
When adding metadata, a unique constraint error appears because the sequence has an incorrect value.

Expected behavior
It is possible to delete all metadata of a harvester at once.
A deviation of index and database should not occur due to the deletion of a large amount of data.

Desktop (please complete the following information):

• GeoNetwork Version: Migrated 4.2.5
• Server Application Tomcat 8

[rime1014]

Does this Draft #8431 fix the above described issue?

[josegar74]

@rime1014 I think so, but the pull request needs some adjustments to work with different types of harvesters.