Currently if the mwacache servers cannot reach the MRO database, it will just fail and exit. We should really handle this- catch the failure (via pyscopg) and just requeue the item and maybe wait some configurable timeout to try again.
Currently if the mwacache servers cannot reach the MRO database, it will just fail and exit.
We should really handle this- catch the failure (via pyscopg) and just requeue the item and maybe wait some configurable timeout to try again.