Zane Bitter 7daa3bdbf3 Ignore errors in purging events ... There is a known issue with purging the resource properties data for events where the same resource properties data is referenced by a resource or event in the backup stack. It's more important that we are able to store a new event than that we are able to purge old events, so catch any exceptions in purging and log an error, rather than allowing the exception to abort the event creation. Also use a transaction to ensure that we roll back the deletion of the events when we encounter an error. This ensures that any resource properties data *not* referenced by the backup stack is not orphaned because of failures deleting other rows. This will allow us to properly purge the database once the issue is fixed. Change-Id: Iefa706f91382f7d1c47e09e8f67a9fad53a9390b Story: #2002643 Task: 22334 (cherry picked from commit 6169ed4660)		2018-06-28 14:14:50 +00:00
..
migrate_repo	Squash database patches	2017-03-14 15:52:46 +08:00
__init__.py	Stub out the database access API	2012-03-29 18:02:16 +02:00
api.py	Ignore errors in purging events	2018-06-28 14:14:50 +00:00
filters.py	replace dict.iteritems() with six.iteritems(dict)	2014-08-06 10:56:09 +08:00
migration.py	Squash database patches	2017-03-14 15:52:46 +08:00
models.py	Data migration enabling storage of resource attributes	2017-02-16 16:08:31 -05:00
types.py	Remove usage of sqlalchemy mutables	2015-06-26 11:02:03 +02:00
utils.py	Replace retrying with tenacity	2016-11-15 04:34:17 +00:00