Cleaning cache takes forever
There's an issue with the pattern for cleaning the cache (_cleancache
in Retriever.py
).
On the test server, name checker actions sometimes keep running for 10 minutes (after which they timeout by gunicorn, we may want to revisit this timeout by the way) and a quick analysis showed that they are in _cleancache
.
It may have to do with the fact that the cache on the test server is 5G while the maximum size is configured to be 800M. But this should still work of course.
Perhaps we should think about an alternative pattern for this, instead of doing administrative tasks during a user request (unrelated to that user request).
Have to investigate further.