[tor-bugs] #13699 [Tor]: Remove rend cache object if not usable anymore

[Tor Bug Tracker & Wiki]

#13699: Remove rend cache object if not usable anymore
-------------------------+--------------------
     Reporter:  dgoulet  |      Owner:
         Type:  defect   |     Status:  closed
     Priority:  normal   |  Milestone:
    Component:  Tor      |    Version:
   Resolution:  invalid  |   Keywords:
Actual Points:           |  Parent ID:
       Points:           |
-------------------------+--------------------
Changes (by dgoulet):

 * status:  new => closed
 * resolution:   => invalid


Comment:

 @special, Yes agree on the fact that we should NOT try to query each IPs
 at every directory fetch.

 > I would rather see a more deliberate decision about when to retry an IP,
 preferably using knowledge about how it failed before (e.g. if the IP was
 reachable or not). What do you think?

 Yes, right now when an "IP failure" is detected, it will be retried 3
 times before being removed from the cache object. If it times out, it will
 not be used anymore and that timed_out flag is reset if the connection to
 the onion address ends at some point (gracefully or not). This all sounds
 reasonable to me.

 After discussion with rransom, this is actually the wanted behaviour
 (original ticket) which is fine. His answer:

 {{{
 No.  If the HS descriptor that a client fetches turns out to match the
 one it already has, it should keep using the (parsed, modified)
 descriptor that it has.  Whether the current descriptor has any
 potentially usable intro points left is irrelevant to that.

 (And it is entirely possible for an HS to publish a descriptor with an
 empty list of intro points, to explicitly indicate to clients that the
 HS is currently offline.)
 }}}

 Thus it makes this ticket invalid.

--
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/13699#comment:4>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online