Roger Dingledine <arma@mit.edu> wrote Tue, 23 Feb 2016 23:16:24 -0500: | On Wed, Feb 24, 2016 at 04:06:46AM +0000, atagar@torproject.org wrote: | > NOTICE: Authorities disagree about the BadExit flag for | > FA00C8CA59ADFAE4859FC5BAD76E68E78C23883E (with flag: moria1, not in | > consensus: gabelmoo, tor26, maatuska, longclaw) | | Looks like that relay is on 5.221.243.105 which is in Iran. | | moria1 has | AuthDirBadExitCC sy,ir | set. So does maatuska. | So this is not surprising on my side. Any ideas from the others? :) Does "not in consensus: <dirauths>" mean that <dirauths> did not vote BadExit for the relay in question? If so, did they vote for the relay at all? Looking at https://consensus-health.torproject.org/ for the 2016-02-24 07:00:00 consensus it seems all five vote badexit for the same 7 relays. On another note, what are y'all using for turning a fingerprint into a public key and ultimately an "identity", as presented in "r" lines?

This was a single one-off alert so I suspect this is a relay that joined the network then immediately disappeared. We get a lot of churn so I should probably modify this DocTor check to only notify once relays have a little uptime.

Change pushed... https://gitweb.torproject.org/doctor.git/commit/?id=9ebead8ccc16955ffa8fd0c2...