[tor-bugs] #20610 [Core Tor/Tor]: Rate limit router_pick_published_address log message

Tor Bug Tracker & Wiki blackhole at torproject.org
Wed Nov 16 15:06:39 UTC 2016


#20610: Rate limit router_pick_published_address log message
--------------------------+------------------------------------
 Reporter:  teor          |          Owner:
     Type:  defect        |         Status:  new
 Priority:  Medium        |      Milestone:  Tor: 0.2.9.x-final
Component:  Core Tor/Tor  |        Version:  Tor: 0.2.8.9
 Severity:  Normal        |     Resolution:
 Keywords:  regression    |  Actual Points:
Parent ID:                |         Points:
 Reviewer:                |        Sponsor:
--------------------------+------------------------------------

Comment (by arma):

 I think the architecture fix that we will want is to recognize that for
 most of the calls to router_pick_published_address(), and especially the
 ones that we call often, what we really mean to ask is "did we figure out
 our address yet?", not "please go figure out our address if we haven't
 done it yet". So if we pass an "only look at the cached variables" option
 to the function, which asks it not to call the heavyweight functions,
 we'll be a lot happier. We'll want to make sure that we call the
 heavyweight functions once near startup. Those steps will make things
 better for the future.

 But first! There is a mystery: It looks like in some circumstances we
 started calling this function a crazy number of times, starting in
 0.2.8.x. And we didn't do that in 0.2.7.x. What changed?

 I would indeed like to get some call graphs or stack traces or debug lines
 the like, for when it happens a whole lot.

 Fortunately, on https://lists.torproject.org/pipermail/tor-
 relays/2016-November/010926.html, pastly said
 {{{
 I run Tor network simulations and see this all the time. I haven't taken
 the time to look into it, and always have assumed it's just a minor
 misconfiguration in the simulator or something that only exists
 _because_ it's a simulation.
 }}}

 so I'm going to take that as a "yes, this crazy number of calls to
 router_pick_published_address is repeatable, and also I can help debug
 it." Is that a fair hope, pastly? :)

--
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/20610#comment:9>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online


More information about the tor-bugs mailing list