The cause is

https://gitweb.torproject.org/tor.git/commit/?id=78e177d622f5f3b24023d04458f...

https://trac.torproject.org/projects/tor/ticket/24803

Would be appreciated if the Tor project published outputs of UpdateFallbackDirs.py job runs used when rebuilding the list. Thus operators who have expended effort to keep their relays eligible will know why when dropped.

Felix zwiebel at quantentunnel.de Sat Dec 15 23:59:03 UTC 2018

Hi everybody

Reading [1] shows the following statistics about the fallback dirs:

Dec 5: Running 116 Not Running 0 Missing 34

Dec 13: Running 155 Not Running 0 Missing 2

Today (Dec 16) shows: Running 151 Not Running 0 Missing 6

I wonder how we can come from 116 running fallbacks to 155 within one week (without a new release):

Then I checked for some fallbacks and found:

The relay with the fp F9246DEF2B653807236DA134F2AEAB103D58ABFE was a fallback acc. fallback_dirs.inc until around 3.2.8rc, then it changed ip. It is still in 2.9.x lts but with the old ip, recommended but actually useless.

Since around Dec 13 it is listed again [1] where it wasn't through 2018.

The same with 8FA37B93397015B2BC5A525C908485260BE9F422.

Are the two relays newly counted as running in [1], where they are useless because of the changed ips? The new ips aren't in any fallback_dirs.inc's and [2] shows no fallback flag. Would thereby [1] look better than it is?

[1] https://consensus-health.torproject.org/consensus-health.html [2] https://metrics.torproject.org/rs.html#details/F9246DEF2B653807236DA134F2AEA...

-- Cheers, Felix