Huge files asset/var/tmp/pdf-text.extract....txt

When importing Assets (PDF) through the REST-API, the system generates huge files in assets/var/tmp/ , named pdf-text-extract-[hash].txt. Every file is writen to 17GB, then a next one is generated.
Only 1.4GB of PDF where imported.
I haven’t found anything related to this files, so anyone know what these files are for and if this is a desired behavoir?

Obviously these files are created by the search index. The maintenance job in our testsystem wasn’t running. After activation of the job the files were processed and deleted.

We are experiencing the same issue since upgrading to v6.6.7. Our maintenance job is running.

pdf-text-extract*.txt in var/tmp getting bigger and bigger until our disk is full.

Any one else with this problem?