Description
We have a Redis cluster running on AWS using their Elasticache server and for reasons I won't go into we periodically change the host that the DB behind it is running on. When we do this, and even with an application (apache/wsgi) rolling restart, we still see failures from the flask-redis plugin that are not recoverable without a reboot. Sounds like something at the connection level, so we probably terminate at that level but a quick reboot does the trick. Has anyone seen this before?
[Thu Nov 09 22:33:06.033764 2017] [:error] [pid 2301] Exception AttributeError: AttributeError("'NoneType' object has no attribute 'error'",) in <bound method Connection.__del__ of Connection<host=**********8.0001.use1.cache.amazonaws.com,port=6379, db=2>> ignored