[tor-consensus-health] Consensus issues

Roger Dingledine arma at mit.edu
Thu Dec 13 20:49:43 UTC 2018


On Thu, Dec 13, 2018 at 05:05:52PM +0000, atagar at torproject.org wrote:
> NOTICE: maatuska had 0 StaleDesc flags in its vote but the consensus had 13
> NOTICE: longclaw had 0 StaleDesc flags in its vote but the consensus had 13
> NOTICE: dizum had 0 StaleDesc flags in its vote but the consensus had 13
> NOTICE: bastet had 0 StaleDesc flags in its vote but the consensus had 13
> NOTICE: gabelmoo had 0 StaleDesc flags in its vote but the consensus had 13
> NOTICE: moria1 had 223 StaleDesc flags in its vote but the consensus had 13
> NOTICE: dannenberg had 0 StaleDesc flags in its vote but the consensus had 13
> NOTICE: Faravahar had 0 StaleDesc flags in its vote but the consensus had 13

Hi Damian,

FYI, I think these particular lines are going to happen always now --
moria1 is running the shiny new code that knows what a StaleDesc flag
is, and the rest of the dir auths aren't. In this case moria1 was voting
StaleDesc about 223 relays, only 13 of which it also voted the Running
flag for (i.e. only 13 of the relays it gave the flag to made it into
the consensus). And I guess since only moria1 expressed an opinion about
the StaleDesc flag, then its opinion is the winner.

--Roger



More information about the tor-consensus-health mailing list