Faster joins: handle total failure to sync state

This issue has been migrated from [#13000](https://github.com/matrix-org/synapse/issues/13000).

---

Currently, if we try every server in the room and are unable to sync state from any of them, we give up, leaving us with a room stuck in "partial state" state, and any C-S requests for state in that room timing out indefinitely.

It's not entirely clear what we should do in this case:
* Giving up isn't the right thing to do if there's a temporary network outage
* Retrying indefinitely is also not the right thing to do if we can reach all homeservers and they all claim they don't have the state we want.

https://github.com/matrix-org/synapse/blob/7c6b2204d143550d81e5bf9612c4e69fe0866b4c/synapse/handlers/federation.py#L1594-L1610