[tor-bugs] #18580 [Tor]: exit relay fails with 'unbound' DNS resolver when lots of requests time-out

[Tor Bug Tracker & Wiki]

#18580: exit relay fails with 'unbound' DNS resolver when lots of requests time-out
----------------------+------------------------------
 Reporter:  Dhalgren  |          Owner:
     Type:  defect    |         Status:  new
 Priority:  Medium    |      Milestone:
Component:  Tor       |        Version:  Tor: 0.2.7.6
 Severity:  Major     |     Resolution:
 Keywords:            |  Actual Points:
Parent ID:            |         Points:
 Reviewer:            |        Sponsor:  None
----------------------+------------------------------

Comment (by Dhalgren):

 Put the exit back to Unbound to see how this works.

 Settled on

    options timeout:5 attempts:1 max-inflight:16384 max-timeouts:1000000
    nameserver 127.0.0.1

 where it turns out that max-timeouts is capped at 255 by eventdns.c.  Will
 create a patch to remove the 255 limit on next Tor daemon update.  Only
 purpose for the "down resolver" state is to shift load to a different
 resolver, but in this situation that's undesirable.  Have exactly one
 local resolver and if it fails an alarm goes off for manual attention.

 I might also create an alarm that triggers when

    unbound-control dump_requestlist

 grows to more than 200 pending requests since that's what was observed
 during the relay failure.  Shouldn't fail now but will be interesting to
 verify that and examine the next DNS potential DOS situation.

--
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/18580#comment:14>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online