collection

SW-455: Merged LOFAR-Releas-3_1 back to trunk so all fixed there are in trunk ready for the 3.2 Release

ROHD-1124: during the rollout of LOFAR-Release-3_1_6 we replaced the radb trigger functions in production with the new functions in SAS/ResourceAssignment/ResourceAssignmentDatabase/radb/sql/add_functions_and_triggers.sql from svn. We forgot that these functions in production had a check to ignore cep4 storage resource, causing problems in scheduling. With this changeset, we (again) ignore cep4 storage claims if they fit. Also added a comment on ideas for faster/better/simpler filling the resource_sage table.

SW-448: merge back from https://svn.astron.nl/LOFAR/branches/SW-448-robinhood-bugfix. Use normal du for now instead of robinhood-du. Easy switch back to rbh-du in october possible. Made queries and scans faster.

SW-448: merge back from https://svn.astron.nl/LOFAR/branches/SW-448-robinhood-bugfix. Use normal du for now instead of robinhood-du. Easy switch back to rbh-du in october possible. Made queries and scans faster.

SW-448: keep copy of current cache_timestamp update estimated path_cache_result in cache, and (re)set the cache_timestamp to the old one, so it will be picked up for a later real du scan

SW-448: update cache entries from bottem to top in dir tree. do fast summation first, and slow du later.

SW-426: merge back from https://svn.astron.nl/LOFAR/branches/SW-426-RT-bugfix. fixed t_schedulers.py tests. Used common testing class for radb tests on a real test-postres-radb instead of faked ones. Solved some more minor bugs in schedulers.

SW-448: added method to do a normal du when robinhood du does not work. For now, let's only use normal du, and switch back on rbh-du when lustre has been updated. Furthermore, added some minor locking/speed improvements.