[tor-bugs] #27436 [Core Tor/Tor]: Rejected v3 descriptor (2018 September edition)

Tor Bug Tracker & Wiki blackhole at torproject.org
Tue Sep 11 12:31:15 UTC 2018


#27436: Rejected v3 descriptor (2018 September edition)
---------------------------------------------+-----------------------------
 Reporter:  asn                              |          Owner:  (none)
     Type:  defect                           |         Status:  new
 Priority:  High                             |      Milestone:  Tor:
                                             |  unspecified
Component:  Core Tor/Tor                     |        Version:
 Severity:  Normal                           |     Resolution:
 Keywords:  tor-hs reliability reachability  |  Actual Points:
Parent ID:                                   |         Points:
 Reviewer:                                   |        Sponsor:
---------------------------------------------+-----------------------------

Comment (by asn):

 OK, I got a few more of those isolated rejections on my HSv3 so I digged
 more. Here are some comments:

 - I asked a friendly relay operator whether he saw any warning logs in
 their relay. They saw nothing, even tho all (?) error paths of
 cache_dir_desc_new()` (called from `hs_cache_store_as_dir()`) lead to a
 `log_warn()`. This might mean that the failure was again a revision
 counter rejection (since that's a `log_info()`), or that I missed an error
 path that doesn't trigger a log_warn, or that the operator missed the log
 path.

 - It does not seem to be a problem with how the rev counters get
 encoded/decoded since descriptors with bigger revision counters got
 accepted just fine, and also all the other HSDirs in that batch accepted
 the descriptor. It's always a single HSDir that's rejecting the
 descriptor.

 - Perhaps we should save descriptors that fail to get uploaded on disk,
 but that does not seem easy given the currrent code architecture.

 In general, because only one descriptor gets rejected out of the whole
 batch, this is not a terrible issue since reachability can still be
 maintained. However, it's definitely a bad bug and it shows there are
 still things we don't understand. We should look more!

--
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/27436#comment:2>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online


More information about the tor-bugs mailing list