Seeder dumping its core after each seeding attempt
Although the mapcache_seed
command crashing even after a successful seeding attempt is mitigated by cache@255f6400 the mapcache_seed command still generates a core dump which is currently roughly 150MB large on PRISM. This means that for each product we seed, such a file is generated in /opt/seeder - working directory.
This is unnecessary and unmaintainable on smaller nodes, where we are limited in disk space.
If we can not fix the original problem in Mapcache or disable the core dumping (I assume we can neither of those), we should at least create a hacky workaround on the python seeder command level deleting the spurious core.x files after each run.