Unreachability of pod propagates through federation
Here's a problem that I encountered while federating with a friend, where I my pod couldn't scan his pod's library because his own pod was federated with a third pod that was now unreachable.
Steps to reproduce
Three actors: pods A, B and C.
Pod B federated once with pod C, now unreachable. Pod B was not following pod C anymore, but domain of pod C was still in the database of B.
I tried to federate with pod B from A.
What happens?
Federation with pod B could not work because of pod C still in the database, which was isolating pod B without him knowing it, because mentions of C were not available to see for pod B admin. Concretely, the scan from A to B times out (even after increasing the timeout to 60) because A tries to reach an URL from C (which is down). This is confirmed by logs from A.
Also deleting pod C from pod B through Django interface was not working because of the size of the request (which failed to complete before nginx timeout), so we had to change that timeout.
What is expected?
I'd expected that scanning B from A does not fail completely if some other actors federated with B are down.
Context
Both A and B are the latest version: 1.2.7, running "bare-metal" (without Docker).