SW-44: removed timing code, and notices. minor performance improvement in get_max_resource_usage_between

SW-44: major simplification of process_new_claim_into_resource_usages and process_old_claim_outof_resource_usages using the new deltas table. Now the new test_20181108_bugfix_resource_usages passes as well, and preliminary results indicate a performance boost as well.

SW-44: moved the integration part of the rebuild_resource_usages* method into a seperate method, for later reusage in handling new/old claim.

SW-44: major refactoring/improvement of rebuild_resource_usages_from_claims_for_resource_of_status using a new resource_usage_delta table. Much simpler code, less exceptions and odd cases.

SW-44: remove obsolete claims once a task ends. This reduces the resource_claim and resource_usage tables significantly, speeding up overall performance.

SW-41: measure and log some more sql procedures to see where the time is spend. report on too slow queries.

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-426: fixes for the new tests in t_radb.py. Main changes: a claim cannot have zero duration anymore. a task which goes to conflict releases its claimed claims

SW-426: found obsolete dangling old file add_triggers.sql. Removed it. It was replaced long ago by add_functions_and_triggers.sql

Task #9893: Fixed tests and added small bug fix for non monitored resources and a fix for tasks being killed in radb before they start.

Task #11232: create_add_virtual_instrument.sql.py to (re)create add_virtual_instrument.sql which contained old and erroneous values.

Task #10902: fixed the test_claims_on_partially_misc_filled_resource by fixing the bug in resource_allocation.get_resource_claimable_capacity_between

Task #10902: fixed and improve (on speed) the rebuild_resource_usages_table_from_claims nethod. Extended the test_resource_usages. Added performance test script which proves the new rebuild is O(N) and about n order of magnitude faster than before.

Task #10902: refactored many selects on resource_usage into method get_resource_usage_at_or_before. Added logging when we detect that a usage_at_start is not found, then the usages table is rebuild for that resource, and then we search again for usage_at_start. Only if not found again, raise exception. This fixes the test for the bug found on production. Still to do: performance increase for rebuild usage table.

Task #10902: added test_reinsert_task test method to prove a bug in production. unfortunately the bug is not proven... so, let's keep this test as reference, and think of a better solution.

Task #11160: fix for the original problem: claims which get 'conflict' status even though they would normally fit when overlapping with others