[tor-bugs] #13287 [Tor]: Investigate mysterious 24-hour lump in hsdir desc fetches

[Tor Bug Tracker & Wiki]

#13287: Investigate mysterious 24-hour lump in hsdir desc fetches
------------------------+------------------------------
     Reporter:  arma    |      Owner:
         Type:  task    |     Status:  new
     Priority:  normal  |  Milestone:  Tor: 0.2.???
    Component:  Tor     |    Version:
   Resolution:          |   Keywords:  SponsorR, tor-hs
Actual Points:          |  Parent ID:
       Points:          |
------------------------+------------------------------

Comment (by dgoulet):

 After a few days, here is a snippet that is consistent across the log on
 my relay:

 {{{
 Nov 20 08:00:00.000 [notice] Heartbeat: 235 descs (29/29 stored, 1/1100
 fetched).
 Nov 20 08:30:00.000 [notice] Heartbeat: 239 descs (24/24 stored, 1/911
 fetched).
 Nov 20 09:00:00.000 [notice] Heartbeat: 238 descs (21/21 stored, 0/917
 fetched).
 }}}

 It's not the huge amount moria shows above but still the fact that in a
 period of 30 minutes, 1100 attempts failed is bothersome.

 Looking at it with nickm, here are possible scenarios for why this is
 happening. (Please add missing scenarios also).

 1) HS doesn't exists
  i) A very popular HS that for some reasons clients are trying to access
 it non stop and it moves around.
  ii) Client that can't find it don't give up?

 2) The HSdir rejected it erroneously?

 3) Clients are asking in the wrong place. (i and ii as above).
  i) Maybe a bug in the tor client that loops?
  ii) Automated client that just retries a single address non stop
 voluntarily (botnet theory)

 4) UNLIKELY: HS enumeration by brute forcing onion address but ouf... that
 is unlikely.

--
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/13287#comment:5>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online