[tor-bugs] #24895 [Core Tor/Tor]: MAX_REND_FAILURES is 1, but we will try three times

Tor Bug Tracker & Wiki blackhole at torproject.org
Tue Jan 16 14:36:10 UTC 2018

#24895: MAX_REND_FAILURES is 1, but we will try three times
 Reporter:  arma                                 |          Owner:  (none)
     Type:  defect                               |         Status:
                                                 |  needs_revision
 Priority:  Medium                               |      Milestone:  Tor:
                                                 |  0.3.3.x-final
Component:  Core Tor/Tor                         |        Version:
 Severity:  Normal                               |     Resolution:
 Keywords:  032-backport, 031-backport,          |  Actual Points:
  030-backport, 029-backport, 025-backport       |
Parent ID:                                       |         Points:
 Reviewer:  dgoulet                              |        Sponsor:

Comment (by dgoulet):

 Datapoint: In months of running v2 and v3 onions, I've seen twice a
 relaunch of a rendezvous point circuit. If the Guard can keep up with the
 service circuit creation, it is something I've rarely seen failing.

 Ok agree that perhaps having two tries to reach the RP is what I think we
 should have in normal circumstances, not only 1 which is the current
 patch. I like the idea of having a consensus parameters so we can adjust
 accordingly depending on the network load.

 But for the current network situation, I think we want to bring it down to
 1 for now because right now 1 million clients introducing would be
 inducing two million circuits by the services they are trying to reach. At
 that scale, I'm ready to call it amplification attack vector.

Ticket URL: <https://trac.torproject.org/projects/tor/ticket/24895#comment:6>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online

More information about the tor-bugs mailing list