[tor-relays] new relay acting strange

[Roger Dingledine]

On Thu, Feb 27, 2014 at 05:24:13AM +0100, Athena Lester wrote:
> Hello,
> This is the first tor relay I have set up, it has been running for going on 4 days. I was waiting for the 3 day lag period I have read about but still no activity, besides the 30 MBs or so per day. It is running on my vps located in the Netherlands. The thing that is strange is when I search for "windywillow" on the atlas it finds two relays, same ip's, same ports. the thing is they both say that the server has been down for the last 3 days when it has been running and recieving a trickle of activity with no adverse log entries. I noticed a day ago that I lost the "Valid" flag. nothing has changed as far as know since it was valid so I'm at a loss.  Here are the links to the two atlas pages.  Any help would be appreciated.
> 
> https://atlas.torproject.org/#details/F51A8E6C811D3C8F2436BE0029E95E67CB0DF3D0
> 
> https://atlas.torproject.org/#details/90B97451822C0881C0652A7DA7430A73BCAF3769

This url:
https://metrics.torproject.org/relay-search.html?search=5.178.64.203
shows that it started out with the F51A... identity key, was in the
consensus for one hour, and then switched keys (perhaps you reinstalled,
or you started running the Tor daemon with a different DataDirectory?),
and then it was in the consensus two times a few days ago with a new key,
and that's it.

Ah ha!

You advertise an ipv6 address:
or-address [2a00:1ca8:37::d85a:e8d5]:29888
which appears to be unreachable.

You are the second person to have this problem -- the directory
authorities are choosing to mark you as non-Running because not all of
your addresses are reachable. I wonder how we can make it more obvious
to the relay operator when this situation occurs?

--Roger