wiki:org/teams/CommunityTeam/Projects/RelayAdvocate

Relay advocate

Fallback list

This can be tracked in the ticket at https://trac.torproject.org/projects/tor/ticket/28794. A certain % of the fallbacks becoming unreachable will trigger the rebuild, but the process itself for opt-ins seems to be run once a year-ish. Its not a complicated process, its largely just adding lines to fallback-scripts repository.

Network monitoring

Relay families

I've been using nusenu's ornetstats to keep track of certain things on the network (https://nusenu.github.io/OrNetStats/ ). Basically, I monitor this for relay families appearing in the E2E correlation position (at the top) and reach out to have them resolve this issue. Also monitoring which families are the largest, making sure they and not disappearing / making sure total consensus weight for a singular family does not get too out of hand.

EOL releases

This page is also useful for getting the list of operators running EOL releases of Tor, and I've been contacting everyone with valid contactinfo every so often on that list to update their relays. You can see all EOL relays at https://nusenu.github.io/OrNetStats/eolrelays

DNS failure

Arthur set up a handy page for monitoring DNS failure rates on exits at https://arthuredelstein.net/exits/. This is another important thing to watch, and to reach out to operators who are experiencing high rates of DNS failure to encourage / assist them in setting up their own local resolvers. The higher the overall failure rate gets, the less happy users will be. Currently its at 2.5%, the highest i've seen it was 10.

We have instructions for setting up unbound in the tor relay manual, which has been enough for most users. However when people want to run something like bind, you may have to find different docs or assist them directly

Currently I am the person relay operators contact if they require a letter from the Tor Project as proof that a certain machine was acting as a relay at a certain time. Shari sent me the process for dealing with these requests along with the letter, I can send these things to gus and/or gus & isa or whoever should see these. I'm not sure if Isa has reviewed the letter. It does not come up super often that these are requested, but when it happens usually there are multiple requests at once. (Some LE contacts 10+ operators about the same thing, they all want a letter, etc..)

So most of the time, its no time investment. But is important to get the letters to these operators in a timely fashion when they are needed, as they usually need to be sent to whoever is requesting the letter fairly quick.

New fast relays

Next is working with people from PIA on relay deployment, I'll send an email making the hand-off there if you're OK with that, Gus (can also wait until you discuss this list with others, of course). Basically there, I'm providing them technical support when their ops people run into trouble, and advising on where new relays would be most impactful for diversity (at their request). So far they have set up 6 high speed exits: https://metrics.torproject.org/rs.html#search/PIA and are continuing to expand. For the most part, the technical support angle is wrapped up as their ops people have done this 6 times now. But sometimes they still have questions.

Relay documentation

Community portal

Next was creating recordings for the upcoming community portal on creating / maintaining relays and bridges. I used asciinema for this, and will be pushing a couple more to my own page / Github that will be usable for this purpose if they are still desired. However future desired platforms will need to be recorded. The current 2 recordings are at https://asciinema.org/~phoul and there are 2 more that will be uploaded probably today on CentOS and configuring Debian as a pluggable transport bridge.

Relay guide to Lektor

Along the same lines, I've been interacting with Nusenu to get the relay manual updated into Lektor, and once that was done my next step was to format our pluggable transport instructions to match. Currently the PT instructions live at https://trac.torproject.org/projects/tor/wiki/doc/PluggableTransports/obfs4proxy He ran into a platform issue that hiro helped him with, and last I heard he's actively converting the manual into Lektor. So I'd expect to have something soon.

Network diversity

Talking with backbone providers / ISPs to accomplish this. I've been collecting contacts and working with ISP operators who are interested in dedicating bandwidth to Tor. There is a large ISP community out there with excess bandwidth and Tor could benefit from a lot of it. I'm happy to make the handoffs with these various people, however I think its important to note that if the general feeling is there wont be time for interacting with them, it might be best to be upfront about that.

Relay meetups

I've got a master list of events I wanted to have relay operator meetups at and contacts for those places, I will send this to you. The two biggest ones on my mind (still a little way out) are DEFCON and the CCC. The CCC in specific *really* wants Tor Project support for this year's meetup, and they have begun reaching out regarding it. Last year there was no meetup at DEFCON, and people were dissapointed by this. There is a large operator community at DEFCON, and there are not many NA conferences with that type of pull.

IRC Meetings

I guess one other thing that isnt in this list is the monthly IRC meetups. I have received messages from a few different people asking if those are continuing. We have had trouble with attendance at the various times the tor-relays list has voted for, so im not sure what the solution is there.

More resources

Last modified 4 months ago Last modified on Aug 30, 2019, 12:39:24 PM